FY1998 Progress – FY1999 Plans


1998 Annual Report –
FY 1998 Progress/FY 1999 Plans

Peter J. Lamb, Director
Randy A. Peppler, Associate Director

The University of Oklahoma (OU) and NOAA established the Cooperative Institute for Mesoscale Meteorological Studies (CIMMS) in 1978. Through mid-1995, CIMMS promoted cooperation and collaboration on problems of mutual interest among research scientists in the NOAA Environmental Research Laboratories (ERL) National Severe Storms Laboratory (NSSL), and faculty, postdoctoral scientists, and students in the School of Meteorology and other academic departments at OU.

The Memorandum of Agreement (MOA) between OU and NOAA that established CIMMS was updated in 1995 to include the National Weather Service (NWS). This expanded the formal OU/NOAA collaboration to the Operational Support Facility (OSF) for the WSR-88D (NEXRAD) Program, the NCEP (National Centers for Environmental Prediction) Storm Prediction Center (SPC), and the NWS Forecast Office, all located on the OU campus in Norman, Oklahoma.

Through CIMMS, OU faculty and NOAA ERL/NWS scientists collaborate on research supported by NOAA programs and laboratories as well as other agencies such as the National Science Foundation (NSF), the U.S. Department of Energy (DOE), the Federal Aviation Administration (FAA), and the National Aeronautics and Space Administration (NASA).

The present 5-year cooperative agreement between OU and NOAA for CIMMS funding took effect on July 1, 1996. Under this agreement, CIMMS concentrates its efforts and resources on the following five principal research themes, the fifth of which is new under the current plan: (1) basic convective and mesoscale research, (2) forecast improvements, (3) climate effects of/controls on mesoscale processes, (4) socioeconomic impacts of mesoscale weather systems and regional-scale climate variations, and (5) Doppler weather radar research and development.

This document describes research progress at CIMMS during fiscal year 1998 (July 1, 1997 through June 30, 1998) and gives research plans for fiscal year 1999 (July 1, 1998 through June 30, 1999), and represents the second annual report written under the present agreement.


Progress – FY 98

Cloud Microphysics

The performance of various liquid water retrieval algorithms for clouds was evaluated based on reflectivity data collected by the ARM southern Great Plains millimeter cloud radar. The evaluation was carried out using observation system simulations. The microphysical data generated by the CIMMS LES model in experiments based on ASTEX and ARM southern Great Plains intensive experimental period observations were used to calculate cloud properties and radar characteristics to verify the retrieval algorithms. The cloud water profiles retrieved from radar reflectivity were compared to those calculated directly from cloud drop spectra. Both non-precipitating and drizzling stratocumulus cloud layers were considered. The retrievals based solely on radar reflectivity data were evaluated against those using a mix of observations from cloud radar and microwave radiometer. It was shown that a combination of cloud radar and microwave radiometer data can significantly increase the accuracy of the liquid water retrieval.

A method was developed for retrieval of cloud optical depth and effective radius for stratiform clouds using the data from the ARM millimeter cloud radar and ground observations of CCN number concentrations. The method was tested using the data from the CIMMS LES model initialized with observations taken during ASTEX and ARM southern Great Plains intensive experimental periods. In addition, the effect of cloud inhomogeneity and variations in cloud base and top were evaluated relative to the performance of various parameterizations of cloud optical depth. The findings are being summarized now for a paper that should be submitted for publication by the end of the 1998.

The CIMMS LES model was combined with a Monte-Carlo radiative transfer model to investigate 3-D radiative transfer in heterogeneous cloud media. The study focused on the role of vertical inhomogeneity in the extinction coefficient, asymmetry parameter, and scattering function. It was found that the area averaged radiation fluxes depended not only on the mean values of optical parameters, but also on their spatial variances. The prediction of the higher statistical moments of optical parameters or the probability distribution functions of cloud microphysical variables may be necessary for more accurate prediction of cloud radiative properties.

Cloud microstructure plays a crucial role in the persistence and fractional coverage of boundary layer clouds that significantly affect regional weather. A cloud drop parameterization was developed in terms of the prognostic variables available in mesoscale models. The CCN concentrations, as well as the boundary layer turbulence intensity, define the number of activated cloud drops. The parameterization was derived using 36 LES numerical experiments in environments characterized by 12 different CCN concentrations (from 25 to 1000/cm³) and three levels of turbulence intensity (characterized by values of in the range from approximately 0.06 to 0.27 m²/s²). The developed parameterization allows a physically based prediction of cloud drop concentration and takes into account the ambient aerosol concentration, as well as the level of turbulence intensity characterized by the mean vertical velocity variance.

Atmospheric Electricity

In February 1998, Oxford University Press published the graduate textbook, The Electrical Nature of Storms, written by Don MacGorman (CIMMS Fellow) and Dave Rust (NSSL). It has been well received by the scientific community.

A new parameterization of lightning for numerical cloud models was tested and refined in collaboration with Jerry Straka of OU and Conrad Ziegler of NSSL. A conference paper (MacGorman et al. 1998) and a journal article (MacGorman et al. 1999) describing the parameterization were submitted.

Parameterizations of electrification processes that were developed for a kinematic cloud model (Ziegler and MacGorman 1994) have been modified and added to the Advanced Regional Prediction System (ARPS) model so that simulation studies of electrification can be performed using that model. Initial test simulations with the electrification modules have succeeded in producing highly electrified supercell storms.

MCS Electrification and Polarimetric Radar Study (MEaPRS)

CIMMS scientists participated in the planning and implementation (including forecasting) of the MCS Electrification and Polarimetric Radar Study (MEaPRS), conducted in central Oklahoma from May 15 through June 15, 1998, and in a small adjunct field program associated with MEaPRS. The two primary objectives of MEaPRS were 1) to investigate mesoscale convective system (MCS) electrification processes, and 2) improve understanding of polarimetric radar measurements. In the MCS electrification component, investigators participated in the mobile ballooning of instruments to record lightning field changes, particle charge, and X-rays. Electric field profiles and airborne radar data were obtained in two mesoscale convective systems. Electric field profiles and multiparameter radar data were also obtained in several isolated severe thunderstorms. These storms and storm systems included several in which the majority of ground flashes lowered positive charge, instead of the more usual negative charge. Acquiring data on such storms addresses the goals of a NSF grant to CIMMS. In some of the storms, the three dimensional location of cloud flashes and ground flashes were mapped by a system developed by New Mexico Institute of Mining and Technology and operated by them in cooperation with NSSL during the field program. These storms will be the subject of considerable research by CIMMS scientists and graduate research assistants during the next few years. In the polarization radar component of MEaPRS, emphasis was placed on collecting high quality microphysical data with which to compare polarimetric measurements made by the NSSL Cimarron radar. NOAA P-3 flights were conducted in seven mesoscale precipitation systems to document the microphysical structure of several MCS stratiform clouds. Complementary data sets consisting of airborne Doppler radar and ground-based polarimetric radar were also collected. Over a three-month period that encompassed MEaPRS, a 2D-video-disdrometer (leased from Joanneum Research of Graz, Austria) collected data that documented drop size distributions. Disdrometer data were collected for ten distinct precipitation events. These included both convective and stratiform precipitation and at least one event that contained large hail. These data were collected approximately 41 km from the Cimarron radar. Initial analyses of these data indicate that polarimetric radar derived rainfall is likely more dependent on drop size distribution than drop shape. CIMMS scientists also documented the performance of the NCEP Eta model during the experiment.

VORTEX Analyses

Analyses of VORTEX field data at NSSL included an ongoing, intensive study of the June 8, 1995, Elmwood, Oklahoma, nontornadic supercell storm. The focus to date has been on a complete four-dimensional analysis, including airborne Doppler radar from the NCAR ELDORA, and the Doppler on Wheels (DOW) radar. Additional data analyzed include those from “Mobile Mesonet” vehicles. Analyses were completed on several time periods prior to and after so-called “tornadogenesis failure.” Preliminary comparisons and contrasts with the tornadic Dimmitt, Texas, supercell of June 2, 1995 indicate that the differences between supercells with tornadogenesis success and tornadogenesis failure may be far subtler than is typically assumed. The Dimmitt tornado appears to fit the conceptual model of tornadogenesis and maintenance as proposed by Rasmussen and Straka.

Analysis of the May 29, 1994, Newcastle, Texas, tornadic storm is nearing completion. This storm will be shown to be a rapidly developing supercell similar in many respects to existing conceptual models of supercells, with tornadogenesis occurring in a manner similar to the conceptual model of Rasmussen and Straka.


The SubVORTEX-RFD (rear flank downdraft) field program (May-June 1998) concluded with excellent data collected on eight nontornadic supercells. The primary goal of this field program was to collect data on the RFD and determine the role that it, inflow, and near-ground circulation and divergence play in both tornadogenesis success and failure (tornadic and nontornadic supercells). Data analysis will begin later this year, as this year’s nontornadic data will be compared to tornadic supercell data obtained during VORTEX.

Storm-Relative Helicity

Several studies related to storm-relative helicity in the environments of tornadic storms have been completed at NSSL. It was found that helicity varies greatly on scales that are not sampled with present observing networks. The causes of this variation are yet to be determined, although small (~ 1 m/s) amplitude hodograph loops and similar variations can lead to large changes in helicity.

Characterization of CAPE

Comparisons at NSSL of convective available potential energy (CAPE) with standard stability indices (used for evaluating the convective potential of the atmosphere) revealed only moderate correlations. This is because indices such as the Lifted Index measure single-level buoyancy while CAPE measures both integration depth and buoyancy. Normalizing CAPE values by the depth over which the integration takes place provides an index (NCAPE) that is independent of the depth and represents a convenient measure of mean parcel buoyancy. This normalization effectively distinguishes between environments with similar CAPE but that exhibit different buoyancy and integration depth. Also, because the vertical distribution of CAPE can play an important role in convective updraft strength, it is advantageous to vertically partition CAPE and NCAPE into multiple layers. NCAPE may provide a more useful indicator of buoyancy in environments in which the depth of free convection is shallow and total CAPE is small. It is suggested that NCAPE computations be used in combination with CAPE for the evaluation of convective potential.

Utilizing Mesoscale Numerical Weather Prediction (NWP) Models to Gain Insight into Forecast Problems

CIMMS scientists in NSSL’s MAG group, in collaboration with other scientists from NSSL and EMC, have facilitated procedures for the utilization of mesoscale models as research tools. A primary motivation for this effort was a desire to use these models for collaborative research projects with forecasters from the SPC. During FY 1998, both the Penn State/NCAR mesoscale model (MM5) and EMC’s MesoEta model were utilized for this purpose. For example, MM5 was used to isolate and understand the important physical processes that led to an unexpected (i.e., unforecasted) changeover from rain to heavy snow during an intense winter storm in February 1998. In addition, both the MM5 and the MesoEta models are being evaluated for their predictive capability in severe weather outbreaks, with a focus on the tornado “super outbreak” of April 3-4, 1974.

Development of a Shallow Convective Parameterization for Meteorological and Air Quality Models

In collaboration with scientists at Penn State and the U.S. EPA, CIMMS scientists at NSSL developed a mass-flux-based parameterization of shallow convective clouds. This parameterization represents the meteorological effects of the complete life cycle of non-precipitating convective clouds. It was fully integrated into MM5 and has been designed to be completely compatible with existing parameterizations of deep, precipitating convection as well as non-convective precipitation processes. Preliminary tests in both 1-D and 3-D frameworks have been very encouraging. In addition, a framework has been established to allow sophisticated air-chemistry models to assimilate and utilize the meteorological variables predicted by the new parameterization.

Balanced and Unbalanced Mesoscale Dynamics

The geostrophic coordinate transformation was applied to the viscous semigeostrophic (SG) Eady wave solution. In the transformed SG space, the inversion of the geostrophic potential vorticity (GPV) becomes a linear problem, so the development of the Eady wave can be clearly interpreted by the interaction between the upper and lower GPV anomalies. The generation of interior GPV anomaly plumes along the upper and lower fronts in the viscous SG solution can be easily interpreted through its analogy to the intrusion of frontal discontinuities into the interior fluid of the inviscid solution after the surface front collapses.

The classic adjoint theory derived for differentiable systems of equations is not applicable to systems with parameterized discontinuities. The classic theory was recently generalized, and the generalized adjoint formulations were further developed to deal with various complex situations in numerical models. The problems can be avoided by introducing coarse-grain tangent linearization and adjoint theory without modifying the traditional discretization, although the coarse-grain gradient check can be performed only for finite perturbations. Generalized adjoint formulations with modified discretization and generalized coarse-grain adjoint theory are derived for a vector system of equations that contains parameterized on/off switches. With vector examples, it is shown that the conventional adjoint minimization may have a convergence problem in multi-dimensional space. The problem can be solved by the generalized adjoint with modified discretization or by the generalized coarse-grain adjoint without modifying the traditional discretization in the forward model.

Adjoint-Method Wind Retrievals from the WSR-88D Network

The ability to retrieve two-dimensional horizontal wind fields from WSR-88D radars has been studied by CIMMS scientists at NSSL. An adjoint-method wind retrieval technique containing predictive equations for reflectivity and radial wind was used. In particular, the potential to retrieve horizontal winds at 500 m from a case involving a non-precipitating outflow boundary was investigated. This research is unique to past single-Doppler radar retrieval studies in that WSR-88D data were used as input to the retrieval technique. Research radars used during field projects, which typically observe the atmosphere at higher resolutions both spatially and temporally than do operational WSR-88D radars, were primarily used here to validate single-Doppler retrieval techniques. In addition, the adjoint-method technique is unique in that a first guess to the retrieval solution is provided through the Velocity-Azimuth Display (VAD) technique. Through the application of the first guess, the adjoint-method retrieval technique has proven to retrieve wind flow with accuracy.

Static Stability and its Evolution Prior to Convective Outbreaks

Two aspects of low- and midlevel tropospheric static stability over the United States were examined by CIMMS scientists at NSSL: the climatological distribution, and the forcing mechanisms responsible for the evolution of static stability prior to and during convective outbreaks. Results show that the Rocky Mountains play an important role in the creation of low stability during most of the year. This result suggests that synoptic flow interacts with and modifies this area of low stability to create an area in the Great Plains where severe convection is possible, when combined with other favorable ingredients. The dominant processes that are responsible for this modification are as yet unclear.

Tornado Vortex Signatures (TVS)

A COMET-sponsored collaborative study of the utility of tornadic vortex signatures (TVS) in very short-term tornado prediction was completed at NSSL. The premise of the study was that guidance provided by a TVS should increase the probability of a successful (verifiable and timely) warning of certain modes of tornadogenesis. By looking at a relatively large sample of tornadoes observed by WSR-88Ds across the country, we have confirmed that such utility depends on TVS “behavior” and also the deleterious effects of radar sampling on TVS representativeness. Behavior is linked to the way in which the tornado develops, namely, (i) formation aloft, then gradual descent to the ground, or (ii) rapid formation either uniformly over a several kilometer vertical depth or very near the ground. Results indicate that tornadoes within convective lines tended to be associated with “non-descending” TVSs, identification of which allowed a mean tornado lead-time of five minutes.

Interpolation of Weather Data to Cartesian Grids

Properties of several techniques used for the interpolation of weather radar data to Cartesian grids were evaluated at NSSL. This was done partly through some theoretical consideration of the properties of the schemes, but was done mostly by empirical testing. In terms of preservation of the phase and amplitude of the input data, predictability of the resultant smoothing and filtering, and relative insensitivity to input data unsteadiness or spatial characteristic, the isotropic Barnes weight function with a constant smoothing parameter appears to be the most desirable of the schemes considered. Modification of the Barnes scheme so that the weight function varies with the data point spacing results in an improved analysis, according to some commonly-used measures of error. Analyses of unsteady input fields, however, suffer from a convolution of data evolution with spatial variations of the weight function. As a consequence, unambiguous assessment of physical evolution is precluded and diagnoses and prognoses based on the analyses become questionable.

NSSL collaborated with NCAR scientists to install the Cartesian Integrated Data Display (CIDD) at NSSL. The system has been installed and has allowed access to national scale data (e.g., radar mosaics, satellite images, and lightning information) and products needed to complete this study. To tackle the RUC-II aspect of this study, NSSL has developed code that reads RUC-II grids, outputs the environmental parameters of interest for specific forecast times, and calculates the mean mesoscale environment for a specified geographical area ahead of a given linear system.

Ensemble Cloud Model Applications to Forecasting Convection

The basic premise used in this NSSL study is that convection on the storm scale is a non-deterministic process, making a probabilistic (ensemble) approach to forecasting convective characteristics more effective than deterministic forecasts (convection usually encompasses a spectrum of characteristics). This approach addresses the conditional probability of storm longevity and type (supercell/non-supercell) given that convection occurs. The COMMAS cloud model was acquired from L. J. Wicker (Texas A&M University) for this investigation. COMMAS was successfully ported to the NSSL/MRAD 10 processor SGI Power Challenge machine, allowing the process of extracting soundings suitable for COMMAS from the MesoEta mesoscale model to be developed and refined. Verification data sets were obtained and processed for 1995 and 1996 using data from the Dallas-Ft. Worth and Memphis WSR-88Ds. A principal component analysis method was developed for characterizing gross storm characteristics based on modeled vertical velocity and modeled reflectivity. This method constitutes a kind of cluster analysis that can result in complete recovery of the original data set from the derived clusters.

Plans – FY 99

Cloud Microphysics

The CIMMS Cloud Physics group is participating in the organization of a special session for the American Geophysical Union meeting scheduled for July 1-4, 1999, in Boston. It is also helping to prepare a special issue of the Journal of Atmospheric and Oceanic Optics in tribute to the late Georgii A. Titov.

Atmospheric Electricity

CIMMS scientists will compare the lightning produced by storm simulations using the new lightning parameterization against three-dimensional maps of observed lightning to test and refine the parameterization.

An analysis of ground flash data, three-dimensional lightning data, electric field profiles, and radar data will be conducted for several storms in which the majority of ground flashes lowered positive charge (positive ground flashes) instead of the more usual negative charge.

CIMMS scientists will also simulate supercell storms to study differences in electrical characteristics between storms that produce many positive ground flashes and storms that produce mostly ground flashes (which lower the usual negative charge to ground). Simulated storm behavior will be compared with observational data sets to evaluate the model results.

NSSL’s WDSS software will be modified to begin ingesting and processing three-dimensional lightning data. This capability will be used to evaluate whether cloud flash data are useful for algorithms that help diagnose severe storm evolution, particularly the onset of severe weather.


Kinematic and microphysical data collected by the NOAA P-3 aircraft during MEaPRS will be combined with balloon-borne electric field meter measurements to refine existing conceptual models of MCS electrical structure and evolution. Additionally, airborne microphysics data collected by the NOAA P-3 will be used to improve knowledge of polarimetric radar measurements in precipitating cloud systems and to investigate both the flux of hydrometeors from the convective line and the vertical microphysical structure of the MCS stratiform cloud. These measurements will allow us to improve understanding of stratiform precipitation processes and MCS water budgets. The 2D-video-disdrometer data collected during MEaPRS will be used to better understand the drop size distributions associated with both convective and stratiform precipitation events. These distributions will then be compared to the NSSL Cimarron radar measurements in an attempt to refine polarimetric rainfall estimation algorithms.

Hurricane Studies

A demonstration project involving many Norman-area scientists is planned for the 1998 hurricane season using DOW radars to document the nature of severe local wind phenomena that occur in conjunction with hurricane landfall.

Low Precipitation Experiment (LPEX)

A multi-agency field program will take place in eastern Colorado in spring 1999. The Low Precipitation Experiment (LPEX) seeks to study low precipitation supercell storms. NSSL and CIMMS will collaborate with NCAR and other universities to develop testable hypotheses and collect field data.

VORTEX Analyses

Additional VORTEX case study work is planned at NSSL. This work includes a focus on tornadic supercells to confirm or refute the findings of the Dimmitt, Texas, case. Also, nontornadic supercells (from SubVORTEX) will be analyzed to compare and contrast with tornadic supercells. Case studies underway or planned involve the Friona (tornadic), Allison (tornadic), Shamrock (nontornadic) and Hanston (tornadic) storms observed in May-June 1995.

Denver Convergence and Vorticity Zone

A small, focused field program will be conducted by CIMMS scientists at NSSL in July 1999 using Mobile Mesonets to study the structure of the Denver Convergence and Vorticity Zone (DCVZ). The goal of the project is to document the existence, scale, and strength of eddies embedded in this zone, which are thought to be the incipient circulations for non-supercell tornadoes when they form. This work will also allow augmentation of Mobile Mesonet data sets of surface boundaries. These data already show that some boundaries collapse to scales of about 100 m or less.

Utilizing Mesoscale NWP Models to Gain Insight into Forecast Problems

CIMMS scientists at NSSL will continue to utilize NWP models as tools to facilitate basic and applied meteorological research. They will continue to cultivate a working relationship with SPC forecasters and to seek guidance from them in defining research topics that are particularly relevant to operational forecasting problems.

Development of a Shallow Convective Parameterization for Meteorological and Air Quality Models

The ability of the new shallow convective parameterization to drive important meteorological processes will be tested at NSSL extensively in a 1-D simulated-Lagrangian framework that has been designed specifically for testing this type of parameterization. Furthermore, the scheme will be tested in the fully 3-D framework of MM5, again from a meteorological perspective. Finally, the parameterization will be evaluated in a sophisticated air quality model, the meteorological component of which is based on MM5.

Balanced and Unbalanced Mesoscale Dynamics

The generation of gravity waves by the unbalanced dynamics associated with mesoscale fronts will be examined. Also, potential applications of the balanced and unbalanced dynamics in mesoscale data assimilation will be explored.

Adjoint-Method Wind Retrievals from the WSR-88D Network

CIMMS scientists at NSSL will add a new divergence constraint to the cost function of the adjoint-method wind retrieval. The Volume Velocity Processing (VVP) algorithm provides divergence observations for this constraint. By employing this constraint, improved convergence/divergence values across the retrieved wind flow should be improved. Wind retrievals produced using the adjoint-method algorithm are unique since WSR-88D data are used as input. Furthermore, retrieved two-dimensional winds fields are produced over larger domains (90 km in diameter) than those in previous wind retrieval studies. Because of the large retrieval areas, results could be used as mesoscale model initial conditions. Much emphasis has been put on using wind retrievals as input to storm-scale models. Only the National Center for Environmental Prediction (NCEP) has attempted to use radar winds in an assimilation scheme for a mesoscale model. However, NCEP assimilates only the radial wind component of the wind rather than a 2-D or 3-D representation of the wind field. In addition, an evaluation of the potential positive impact of using the radial wind data in the assimilation process has not yet been assessed. Therefore, the question of the impact of wind retrievals on the improvement of simulating mesoscale features within mesoscale models remains a challenging problem to be answered. This particular problem will be addressed in the upcoming year. A paper with the above title will be submitted to the Journal of Applied Meteorology during fall 1998.

Static Stability and Its Evolution Prior to Convective Outbreaks

During the next year, the mechanisms that cause static stability to change will be examined at NSSL for a severe convective outbreak that produced a significant amount of tornadoes across the southeastern U.S. Preliminary results suggest that horizontal advection of lower stability values is the dominant process for decreasing stability; other less significant processes include vertical advection of stability, vertical stretching, vertical motion, differential diabatic heating, and differential thermal advection.

The Use and Misuse of Conditional Symmetric Instability

A commonly employed explanation for single- and multiple-banded clouds and precipitation in the extratropics is slantwise convection due to the release of moist symmetric instability (MSI), of which one type is conditional symmetric instability (CSI). Research at NSSL will review CSI with the intent of synthesizing previous observational and modeling studies. It is our contention that CSI as a diagnostic tool to assess slantwise convection has been, and continues to be, misused and overused. Drawing parallels to an ingredients-based methodology to forecasting deep moist convection, which requires the simultaneous presence of instability, moisture, and lift, some of the misapplications of CSI can be clarified. Many of these pitfalls have been noted by earlier authors, but are, nevertheless, often understated, misinterpreted, or neglected by later researchers and forecasters. Topics include the evaluation of the potential for MSI, the relationship between frontogenesis and MSI, the coexistence of moist gravitational instability and MSI, the nature of banding associated with slantwise convection, and the diagnosis of slantwise convection using mesoscale numerical models. The review will conclude with directions for future observational, diagnostic, and theoretical investigation.

The Structure and Evolution of Cyclones and Fronts over the Western United States

This research is part of a collaborative effort with between OU and NSSL. The evolution of the midlatitude cyclone of 12-14 December 1988 will be described from its appearance over the eastern Pacific Ocean, to landfall on the coast of western North America, to the progression of a lower-tropospheric cold front across the Pacific Northwest, and to its redevelopment on the lee side of the Rocky Mountains. This research, though primarily observation-based using the standard observational network, will also employ the nonhydrostatic MM5 to assist in the analysis. In this way, an understanding should be gained of the evolution of a cyclone and its associated fronts over the western United States using the same tools available to operational meteorologists in real time. Each stage in the evolution of this event addresses scientific issues relevant to both operational and research meteorologists. For example, in the landfalling stage, the structure of the mid- and upper-tropospheric baroclinic zone undergoes an evolution previously documented only for upper-level baroclinic zones in northwesterly flow. As the system moves through the Pacific Northwest, a nonfrontal pressure trough and wind shift occur at the surface, apparently associated with the upper-level baroclinic zone. These observations are relevant to the issues of surface frontal analysis addressed by Sanders and Doswell (1995) and Sanders (1998). As the system moves through the Rockies, frontogenesis results on a variety of scales (mountain, meso, and synoptic), leading to the development of a warm front in Montana. The movement of the cold frontal zone will also described.


A climatological and theoretical/modeling study of tornadogenesis within convective lines will be pursued at NSSL.

Winter Precipitation

Data collected from NSSL’s dual-polarization Cimarron radar, in addition to other data sets, will be used to investigate the microphysics and mesoscale dynamics of winter precipitation events.

Ensemble Cloud Model Applications to Forecasting Convection

Approximately 500 to 700 cloud model runs (depending upon how many verification days are available and how much MesoEta output can be obtained) will be made at NSSL on soundings for which verification data are in-hand. A determination will be made whether these model runs can reproduce the observed distribution of cell lifetimes using measure such as vertical velocity or reflectivity above appropriate thresholds. Also, for the limited number of verification data sets containing supercell type storms, a determination will be made whether this combination of MesoEta soundings and COMMAS cloud model can create supercells in and around the areas where they were observed. Establishing that these two characteristics can be reproduced will allow the possibility of numerical, objective ensemble-derived probability forecasts for the spectrum of convective weather expected within a given area.


Progress – FY 98

OSF-OTB/OCCE Distance Learning

The NWS requires validated objective assurance that the shift from on-site to distance-learning delivery systems for training has not degraded forecast and warning skills. This vital issue is being addressed through collaboration between CIMMS, OSF-Operations Training Branch (OTB), and the OU Center for Distance Education (part of the OU College of Continuing Education – OCCE). The focus of the project, which recently started, has been threefold:

1. Are forecasters (students) in distance-learning courses meeting training objectives at the same high rate as forecasters (students) trained in past residential courses?

2. Are NWS supervisors satisfied with the results of the new distance-learning delivery system for training?

3. Are all the components of the new delivery system functioning with equal effectiveness? If not, what recommendations might be made for enhancement and further investigation?

Initial results, gathered from multiple response instruments and metrics, indicate that the distance-learning delivery system is producing high quality training, measurably equivalent to the past system. The high quality of the training continues to allow forecast improvements to be made at local forecast offices.

SPC Forecast Verification

During the past year, CIMMS staff at SPC have started working toward redeveloping SPC’s forecast verification program. When the SPC moved from Kansas City to Norman two years ago, SPC’s verification program was discontinued and was in need of replacement. The old methods of verification were applied to SPC products for the years 1970 through 1997 so that a climatological record of forecast verification was available through which to judge forecaster performance. Most of the verification work to date has been focused on Convective Watches (severe thunderstorms and tornadoes), which is one of the primary public services that the SPC offers.

Local Software Development within the SPC

Software development has started at SPC that allows the forecaster to interrogate non-synoptic upper air wind data sources. These data sources include VAD wind profiles (VWPs) and wind profiler data, as well as traditional radiosonde data. Owing to the lack of coverage of the twice-per-day radiosonde network in the U.S., the SPC depends heavily on the other asynoptic sources of upper air wind data for the issuance of the Convective Watch and Convective Outlook products that provide national guidance to forecast offices on severe weather potential. A graphical user interface (GUI) is being developed that allows forecasters to use data to their fullest extent. Updating and augmenting such information on the SPC web site has improved services to SPC customers.

Improving Operational Numerical Prediction of Convective Weather Systems

An experimental version of the Environmental Modeling Center’s (EMC’s) Eta model has been run on a daily basis for most of FY 1998 at NSSL, in parallel with the operational integration of the model at EMC. Model output has been disseminated by means of a web site (http://vicksburg.nssl.noaa.gov/etakf/) and incorporated into the model diagnostic package available on the SPC’s NAWIPS workstations. In addition, 24-h precipitation totals are available for comparison with observations and predictions from other EMC models at another web site (http://sgi62.wwb.noaa.gov: 8080/verf/pcpgifs.html). The primary distinguishing characteristic of NSSL’s configuration of the model is the inclusion of the Kain-Fritsch (KF) convective parameterization. This parameterization is designed to enhance the capability of the model to generate mesoscale structures that are commensurate with its spatial resolution. Preliminary assessments of the experimental model’s performance have been conducted in collaboration with forecasters from the SPC and scientists from EMC. Operational forecasters have indicated that, in a qualitative sense, the KF scheme introduces considerable additional predictive skill related to the timing, location, intensity, and evolution of deep convection in many types of convective environments. However, EMC scientists have noted that the operational version of the Eta model has generally outscored the experimental version in traditional measures of skill in quantitative precipitation forecasting.

Hazardous Winter Weather Climatologies

CIMMS scientists at NSSL have participated in a project designed to provide forecasters at the SPC with climatological information on hazardous winter weather phenomena over the United States. It began three years ago because of the need to analyze and publish climatological data related to these hazardous winter events. During the past year, meteorological data has been analyzed to document the climatology of freezing rain and reports of thunder with subfreezing temperatures. The results of these studies indicate that freezing rain events are rare mesoscale events that occur primarily over the eastern United States from December-March. Four areas within the United States have received a relatively large number of freezing hours each year: eastern Oregon/Washington, western North Carolina, northwestern Pennsylvania, and central Missouri. Northwestern Pennsylvania, near Bradford, has received the largest number of freezing rain hours. Synoptic conditions that are associated with these events include surface temperatures near freezing, midtropospheric shortwave troughs, and strong low-level (~850-800 hPa) warm advection. These events typically are oriented west-east north of warm fronts and last less than one hour. Very few of these reports are accompanied by thunder; the central United States experienced the most thunder events that occurred with subfreezing temperatures.

Warning Decision Support System (WDSS)

For five years, CIMMS and NSSL personnel have been developing and testing the Warning Decision Support System (WDSS). The WDSS has been tested at 17 WFOs around the country since 1998. Many of the concepts developed in the WDSS have received very favorable comments from operational forecasters at WFOs. During 1998, CIMMS and NSSL personnel worked very closely with the Techniques Development Laboratory (TDL) to implement WDSS functionality in the NWS Advanced Weather Interactive Processing System (AWIPS). Many of the most well received features of WDSS have been implemented into AWIPS and are slated for release during 1999.

Areal Mean Basin Estimated Rainfall Algorithm

The Areal Mean Basin Estimated Rainfall (AMBER) algorithm was originally developed by meteorologists at the Pittsburgh WFO. During 1998, the algorithm was implemented at NSSL to run in real-time as part of WDSS. The implementation included the development and deployment of a display utility that provides Flash Flood guidance information to meteorologists. This implementation marks the first Flash Flood decision support tool to implemented for real-time operational feedback within the NWS. The feedback obtained from the testing and evaluation of the algorithm and display will be utilized in implementing this algorithm or similar applications in AWIPS.

Satellite-Based Anvil Tracking Algorithm

During 1998, CIMMS personnel at NSSL began developing a Satellite-based Anvil Tracking Algorithm. The objective of the algorithm is to detect and track the cold top anvil features that are often associated with severe storms. The algorithm will lay a foundation for further association of other known severe weather signatures (such as enhanced V-notch and warm wake features) with detected storms in satellite data. The development progressed to a state where the algorithm could be tested in a real-time operational setting. The algorithm ran operationally in the Sterling, Virginia, WFO during August and September 1998.

System for Convection Analysis and Nowcasting (SCAN) Implementation and Testing

For the second year, the System for Convection Analysis and Nowcasting (SCAN) was tested in a real-time operational setting at the Sterling, Virginia, WFO. SCAN is being developed and designed as a future AWIPS to provide short-term guidance to forecasters on the probability of severe weather. NSSL, NCAR, and TDL are undertaking the development of the system. The 1998 test will run from June 1 – October 30 and will provide feedback on the utility of new severe weather applications for operational warning situations.

Reflectivity-Derived Thunderstorm Parameters Applied to Storm Longevity Forecasting

The FAA requires storm evolution forecasts to aid Air Traffic Control decision-making regarding flight departures and landings. In 1997, CIMMS scientists at NSSL investigated the value of NEXRAD storm characteristics for storm longevity forecasts. Sixteen characteristics, as determined by the Storm Cell Identification and Tracking (SCIT) algorithm, the Hail Detection Algorithm (HDA) and storm longevity were determined using univariate and multiple linear regression analyses. The study included 879 storms (4990 volume scans) that formed over the Memphis, Tennessee, vicinity during 15 late spring and summer days. The purpose of this study is to determine if any of these relationships are strong enough to warrant the inclusion of one or more storm characteristics in the development of storm longevity forecasts. Both univariate and multivariate linear regression analyses of all storm characteristics and remaining storm lifetime showed that remaining lifetime discrimination based on these storm characteristics is relatively poor. The storm maximum reflectivity was most strongly related to remaining lifetime (correlation of 0.36), while the combined storm characteristics were slightly more related to remaining lifetime (0.43).

Velocity-Derived Thunderstorm Parameters Applied to Storm Longevity Forecasting

The prediction using NEXRAD storm cell parameters study came to a close this past January. The final phase of this study investigated the relationships between storm longevity and storm rotation parameters. The data set contained 648 storms (5175 volume scans) associated with at least one MDA circulation during its life. Two attributes were used to categorize storm rotation strength: the Mesocyclone Strength Index (MSI) and storm rank (rank >= 5 is a mesocyclone). Discrete probability density functions (PDFs) were created for the following categories: 1) No circulation, 2) circulation present, 3) MSI >= 2000, and 4) rank >= 5 (mesocyclone detected). These PDFs showed the following:


% Dissipating < 30-min

% Dissipating >= 30-min

Mesocyclone (N=438)



MSI >= 2000 (N=909)



Circulation (N=2536)



No Circulation (N=1292)



The remaining lifetime distributions for both a mesocyclone and a MSI >= 2000 were very similar, but MSI >= 2000 occurred more frequently than did mesocyclone detection. What the MSI and rank distributions tell us is that, given an MSI >= 2000 or a rank >= 5 associated with a SCIT identified storm at some time t, about 70% of the time the storm will last 30 minutes or longer.

Data Collection and Data Use by Forecasters and Public Service Agencies

Maintenance and operation of the Oklahoma Mesonetwork continued at the Oklahoma Climatological Survey. Reliable data services included greater than 98% real-time data availability and over 99.9% availability of research-quality archived data. NWS data users during FY1998 included Weather Forecast Offices in Norman, Tulsa, Amarillo, Dodge City, Wichita, and Shreveport, the Storm Prediction Center, and the Arkansas-Red River Basin River Forecast Center in Tulsa. Mesonet data were used frequently in short term nowcasts as evidenced by numerous references to the Mesonet in various “State Forecast Discussions”. In addition, case studies using the data served as important training tools on which to improve NWS operations. The data will continue to impact the NWS as it continues its modernization program, especially in the areas of hydrometeorology and radar rainfall calibrations.

The OK-FIRST Project (OKlahoma’s First-response Information Resource System using Telecommunications) is nearing completion with its present funding. It is an initiative by the Oklahoma Climatological Survey to improve access to and the use of weather and environmental information by public safety agencies. More environmental information is generated now than ever before, but public safety agencies such as emergency management, fire, and police have had inadequate telecommunication infrastructures to fully access and exploit such information to make informed environmental-based decisions. Sufficient training on how to use such information has also been lacking. OK-FIRST is supported partially by the U.S. Department of Commerce to provide these services to such agencies, and it has been a collaborative project. A first step in the project included an OCS partnership with Unisys Weather Information Services to allow cost-effective redistribution of WSR-88D data. The State of Oklahoma became involved as an Internet provider through its new “OneNet” system. To date, several training sessions involving over 50 public safety agency participants have been held at OU.

Validation of Hydrologic Models

The validation of hydrologic models that will be developed for GCIP (GEWEX Continental International Project; GEWEX = Global Energy and Water Cycle Experiment of the World Climate Research Programme) will depend, to a significant degree, on the accuracy of precipitation data, since precipitation is the largest component of the water cycle. Two means of measuring precipitation are raingauges and radar, and both are susceptible to errors in observation. A three-year study was undertaken to assess the accuracy (bias and error variance) of hourly rainfall as estimated by WSR-88D NEXRAD radars. The reference data sets for determining this accuracy are from a network of raingauges in Oklahoma. An important goal of the research is to determine the effect of wind on undercatch for raingauges used in building the reference data set and the effect of sampling error on the validation process. A raingauge testbed was constructed to investigate the undercatch problem. Part of the effort to facilitate this research included establishment of the Environmental Verification and Analysis Center (EVAC), which collaborates closely with GCIP in developing and applying statistical averaging techniques designed to minimize the error arising from incomplete spatial sampling and field inhomogeneities. EVAC has allowed scientists the opportunity to obtain assessments of quality concerning the accuracy of remotely sensed variables and model output. It has also allowed for two-way data exchange to enhance the quality of the GCIP database and access to data by research scientists.

Plans – FY 99

OSF-OTB/OCCE Distance Learning

Project plans include item analysis of archival data from past residential Operations Courses, correlation and factor analysis of past residential test scores with current distance delivery scores, validation of instructional methods, review and synthesis of client satisfaction surveys, analysis of curriculum objectives with outcomes, and site visits to representative field offices including in-depth interviews with students and supervisors. This work will continue to introduce, examine, and analyze present and future technologies vis-à-vis technology transfer with the goal of meeting unfulfilled needs of the NEXRAD program. The final report on the collaborative project to evaluate the effectiveness of OSF/OTB distance learning is due in late 1998. It will provide a list of recommendations for further investigation of report findings and for enhancements to distance-learning training methods. With the continued goal of forecast improvements through better distance-learning training, additional investigation and development of enhancements will likely occur during FY 1999.

Objective Limits of Forecasting Skill for Rare Events

NSSL and SPC are collaborating on a project to determine the objective limits of forecast skill when forecasting rare events. This project aims to find reasonable daily severe weather forecast statistics against which forecasters can be judged. This work is attempting to account for the fact that severe weather occurs, in general, over a very small fraction of the area over which the SPC issues guidance, and thus, determine how verification statistics can accurately reflect “practically perfect” forecasts.

Local Probabilities of Severe Weather

Research will continue on determining local probabilities for severe weather. In the next decade, the SPC will likely begin issuing probabilistic guidance for forecast products rather than using arbitrary thresholds for severe weather occurrence or intensity. Preliminary work is being conducted to look at the feasibility of developing a national climatology of severe weather occurrence as it relates to SPC Convective Watches and Convective Outlooks. These probabilities, once developed, would give NWS field forecasters guidance on the likelihood of severe weather in their area based on SPC decisions to include their area in certain categories of Convective Watches or Convective Outlooks.

Local Software Development at the SPC

Owing to uncertainties and limitations contained within numerical model data, SPC forecasters continue to rely heavily on observational data when making short term decisions related to Convective Watches. A need remains to continue the development of software to allow SPC forecasters to interrogate all available data to their fullest extent.

SPC Forecast Verification

Work will continue on SPC forecast verification. This work will include the verification of SPC Convective Outlooks and Convective Watches. In addition, work will begin to implement the verification schemes in near real-time so that forecasters can judge their own performance as quickly as possible. Once a verification scheme is operational, work will begin on a next generation of forecast verification schemes in collaboration with researchers at NSSL. This work will begin to address some of the problems with current verification schemes as well as developing schemes for the verification of future SPC probabilistic forecasts.

Climatology of SPC Products

Interactions with NWS forecasters have revealed that they are interested in obtaining climatologies of SPC Convective Watches so that they can have some idea of when and where to expect the SPC to issue watches. Currently, SPC Convective Watches verify with severe weather occurrences approximately 90% of the time. By having a Watch climatology, NWS forecasters may be able to adjust staffing levels and perform other advanced planning in order to provide better services to the communities for which they are responsible.

Climatology of Supercells

An examination was conducted at NSSL of all 1992 upper air soundings that had nonzero CAPE. Soundings were classified as representative of supercells producing significant tornadoes, supercells not producing significant tornadoes, or non-supercells. SELS log and lightning strike data were used to identify events. Based on this limited sample size, some interesting characteristics have been observed. The original one-year study will be expanded to 10 years to provide greater statistical significance for the results and to evaluate year-to-year variability.

Improving Operational Numerical Prediction of Convective Weather Systems

The experimental version of the Eta model will continue to be run at NSSL. CIMMS scientists in NSSL’s Mesoscale Applications Group (MAG) will work closely with forecasters at the SPC with two main goals in regard to this experimental model. The first will be to continue to educate the forecasters on the basic operating principles of numerical weather prediction models, particularly the elements of these models that most significantly affect the prediction of deep, potentially severe convection. The second will be to utilize the keen insight and daily experience of the forecasters to design and implement alternatives to traditional measures of numerical weather prediction skill. In particular, the efforts related to operational guidance from the experimental version of the Eta model will focus on the development of model-evaluation techniques that measure a model’s reliability in predicting deep convection and its associated mesoscale features. This collaboration will include active participation in SPC forecast shifts by MAG group scientists and active participation in model evaluation by SPC forecasters.

Hazardous Winter Weather Climatologies

Much of the preliminary work with this project should be completed by the summer of 1999. The results of this project will be submitted to an American Meteorological Society publication sometime during the next reporting period. New research will concentrate on learning more about the dynamics associated with these events.

Warning Decision Support System (WDSS)

Many of the most well received features of WDSS have been implemented into AWIPS and are slated for release during 1999. Plans for the remainder of 1998 and 1999 also include extending the new AWIPS WDSS functionality to include more WDSS features.

AMBER Algorithm

The AMBER algorithm will be tested in the Sterling, Virginia, WFO during the remainder of 1998 and at the Tulsa, Oklahoma, WFO during late 1998 and 1999.

Satellite-Based Anvil Tracking Algorithm

Plans call for the Satellite-based Anvil Tracking Algorithm and associated displays to be run during the fall and winter months of 1998 and an enhanced (more automated) algorithm to be implemented during the spring of 1999.

SCAN Implementation and Testing

Plans for SCAN during 1999 include additional testing (with new applications) at a site yet to be determined.

Demise of Organized Convective Systems

NSSL has started a study that has as its objective to determine methods for forecasting the demise of organized convective systems one to three hours in advance. The study will first focus on forecasting the dissipation of line storms. Previous MCS studies have shown that the mesoscale environment ahead of an organized convective system plays an important role in the sustenance or demise of the convective system. Therefore, we hypothesize that mesoscale environmental parameters such as CAPE, CIN, vertical velocity (700 hPa), DCAPE and others, forecast out ahead of a convective system, may be useful for predicting convective demise. We will test this hypothesis by analyzing the relationship between the RUC-II predicted temporal evolution of the storm-relative mesoscale environment ahead of a few linear convective lines, and the system’s observed evolution.

Data Collection and Data Use by Forecasters and Public Service Agencies

OK-FIRST progress has been greater than anticipated. All feedback received from public safety participants and funding agencies has been extremely positive. Work will continue through at least September 30, 1998.


Progress – FY 98

‘El Moubarak’ and Moroccan Precipitation

Over the last four years, there has been a strong collaborative effort between CIMMS and the Kingdom of Morocco to increase our understanding of the interannual-to-decadal variability of Moroccan winter precipitation within the North Atlantic climate system, as well as within the global climate system as it relates to the late rainy season (tropical Pacific sea surface temperatures). This understanding is being used to develop a seasonal precipitation capability for Morocco. This project was motivated by the predominance of extremely poor Moroccan winter precipitation since the late 1970’s. During the 1997-98 Moroccan precipitation season, CIMMS issued an Experimental Prediction Statement plus three updates/verifications to officials in the Moroccan government. These long-term predictions were based exclusively on CIMMS research findings over the last four years. During the first half of FY 1997, five presentations were made at international meetings to present our findings. These included: Dr. Peter Lamb presented Prediction and Decadal-Scale Ocean Climate at the International Council of the Exploration of the Sea (ICES) Workshop in Copenhagen, Denmark and On the North Atlantic Oscillation and the Seasonal Prediction of Moroccan Precipitation at the First Regional Training Course on Practical Applications of Seasonal-to-Interannual Climate Prediction to Decision-Making in Agriculture and Water Resources Management in Africa in Niamey, Niger; Dr. Neil Ward gave an invited presentation on Blending of Modeling and Statistical Approaches to the WMO Workshop on Dynamical Extended Range Forecasting in Toulouse, France; Diane Portis presented a poster at the 22nd Annual Climatic Diagnostics Workshop detailing the research basis for, the nature of, and verification of our Prediction Statements; and Mostafa El Hamly, a Moroccan engineer from the Direction de la Météorologie Nationale in Casablanca, gave a paper to the Annual Meeting of the AGU on the oceanic and climatic signature of the North Atlantic Oscillation. During the spring of 1998, our group co-authored a book chapter, North African Climate Variability, which will appear in “Beyond El Niño: Decadal Variability in the Climate System” to be published by Springer Verlag. This book chapter features research on the climate variability of the Sahel and Maghreb.

During FY 1997, a new phase of the Moroccan project was initiated to study the relationship between weather systems and Moroccan precipitation patterns within the North Atlantic climate system. In the past, research has been limited to analysis of monthly fields of the North Atlantic Oscillation (NAO) index and Moroccan precipitation. With the introduction of the analysis of weather systems into the research, higher resolution timescales are needed. Daily precipitation data from 25 Moroccan stations during 1979 have been obtained and have been quantified into rainfall events. For the documentation of the weather systems, a cyclone-tracking algorithm was obtained from Dr. Serreze of Colorado State University and it has tracked the intensity and the path of weather systems in the North Atlantic since 1958 using four times per day NCEP/NCAR reanalysis data.

Regional Climate Prediction in Africa

CIMMS scientists participated in regional climate outlook forums in West Africa, East Africa, and South Africa. Training at ACMAD in Niger prior to the West African forum led to the national meteorological service there developing its own national seasonal prediction schemes using sea surface temperature predictors.

CIMMS scientists undertook collaborative work with the South African Drought Monitoring Center to develop the first system for verifying regional climate outlook forum seasonal outlook maps.

As part of the World Bank funded “Environmental Management Project” in Malawi, CIMMS scientists made two visits to the Malawi Meteorological Service. Reports were written to assist the development of climate information and prediction services for the private and public sectors, including recommendations on the technical infrastructure and training required.

Predictability of East African Rains

Studies into the predictability of the East African October-December rains were continued. A particular focus was the interaction between the synoptic, intraseasonal, and seasonal time scales. This provided a testbed application for the NCEP reanalysis products. Initial studies used five-day means of 850 hPa and 200 hPa wind anomalies. Major five-day wet spells in East Africa were identified using five-day outgoing longwave radiation anomalies over 1979-94. A sequence of global tropical development was found to precede the rainfall events as follows:

  • 10-15 day lead: A burst of near-surface divergence occurred in the western Pacific with 850 kPa easterly anomalies across the Indian Ocean and westerly anomalies across the central Pacific
  • 5-10 day lead: Westerly Pacific anomalies transferred to the tropical Atlantic
  • During the event: Westerly anomalies penetrated across equatorial Africa and into East Africa.

These findings were presented at the Kenyan Meteorological Society Workshop in September 1998 and they generated considerable interest among the national meteorological services in the East African region. Many of the regional findings confirmed “local working knowledge”, while others (like the Pacific and tropical Atlantic 850 kPa westerlies) appeared completely unexpected. These new findings have the potential to assist with early warnings of wet spells with 5-10 day lead times.

Investigations of Climate Variations and Trends Associated with Tropical Variability

CIMMS and OU scientists are investigating linkages between decadal scale variability in sea surface temperature anomaly fields in the Pacific Ocean and the concomitant jump in tropical geopotential heights observed in the mid-1970s. A dissertation is being prepared by Xiaofeng Gong.

Application of WSR-88D Data Toward Developing Four-Dimensional Wind and Cloud Fields

WSR-88D (Level II) radar data provide an additional external data set for research directed toward achieving the scientific goals of the ARM Program. This project involves efforts in two unique but related areas. The first area entails the creation of vertical profiles of the horizontal wind and divergence at radar locations within the southern Great Plains ARM site. These profiles can be obtained by using pre-existing algorithms supplied by the Stormscale Research and Applications Division of NSSL in cooperation with the OSF. Preliminary work, conducted at CIMMS within the last year, has demonstrated that the radar-derived velocities and divergence signatures can be used to verify the large-scale vertical motion fields provided in the ARM Single-Column Model (SCM) data sets (or from models such as that of the ECMWF). The second area of interest involves the integration of the radar data into 4-D analyses of the cloud field within the SCM box. Algorithms, developed at the Center for the Analysis and Prediction of Storms at OU, which integrate a variety of observation sources (e.g., radar, satellite, and surface data), will be used to generate cloud fields and associated products. Evaluation of the advective tendencies of hydrometeors (which are not currently well estimated) is an anticipated output of this work. The performance of the cloud analysis package will, in part, be evaluated via an intercomparison of ARM observations (e.g., millimeter cloud radar) with the analysis cloud fields. Both research thrusts are expected to contribute to the development and testing of SCMs. This work involves collaboration with scientists at the University of Utah and the U.S. Department of Energy.

Tropical Western Pacific Solar Radiation

The objective of this project was to build an historical database of daily solar radiation values for all of the reporting stations in the tropical western Pacific (100°E-170°E, 10°S-10°N). There were 212 possible observing stations in this region. The daily solar radiation values were generated using a sophisticated semi-physical model which has been under development since the mid-1970’s (Atwater and Ball, 1978; Meyers and Dale, 1983). This model estimates the instantaneous radiant flux density received at the Earth’s surface, which is a product of the incoming solar radiation at a location on the Earth’s surface and the attenuation of this radiation from clouds, aerosols, Rayleigh scattering, and absorption by water vapor and permanent gases. This unique tropical dataset, which gives long-term representative information of the Sun’s radiant heat energy, should be of wide utility to the climate, agricultural, hydrological and engineering communities. This project has been completed.

Land-Surface Interactions

Land-atmosphere interactions are central to the natural environment, involving and affecting individual weather systems, regional climate, the hydrological cycle, soil and vegetation status, and agricultural production. This importance is manifest in several long-term, international, biological and geophysical programs, including GEWEX. The first major GEWEX activity, GCIP, is focusing on the land-atmosphere interactions of the greater Mississippi River basin during the second half of this decade. A land-atmosphere interaction suite of pervasive environmental importance involves the classical issue of the relative contributions to regional precipitation of locally evapotranspired (i.e., recycled) moisture versus externally advected atmospheric water vapor. Using a new but simple formulation, we obtained the first comprehensive estimation of the intraseasonal and interannual variability of those moisture sources for the growing season precipitation of arguably the World’s most productive, largely unirrigated, agricultural region — the Corn Belt and surrounding areas of the Midwestern United States — which also occupies ~35% of the GCIP domain. Consistent with its GCIP co-location, this region is considered representative of the mid-latitude, mid-continent, land-atmosphere interactions that are vital for global water resources and hence food production. For four highly contrasting growing seasons, we found the contribution of the locally evapotranspired moisture to this precipitation to be relatively small and remarkably consistent (largely 19-24%) on a monthly and seasonal mean basis, despite large precipitation and crop yield variations, and to decrease markedly (from a 28% to 15% average) with increasing precipitation on a daily basis. Our approach and results yielded considerable physical insight into the complex land-atmosphere interactions involved, including into plant behavior and the apparent paradox between the above monthly/seasonal and daily time-scale results. The moisture budget components and related parameters were evaluated for a large area (about 106 km²) in the Midwestern United States for all 24-hour (12-12 UT) periods during the highly (contrasting) May-August periods of 1975, 1976, 1979, and 1988. Relationships among the budget components were obtained by first stratifying them in different ways, and then by using linear correlation and cross-spectral analysis. The results showed that the calculation of evaporation as a residual of the moisture budget equation yielded values close to the (few) existing observations, especially for periods on the order of one month or longer. The evaporation showed a clear bimodal distribution with respect to precipitation, with high evaporation associated with low and high precipitation amounts, and with a minimum of 3.1 mm d-1 for a precipitation rate of about 4-5 mm d-1. The interannual and intraseasonal variation in precipitation was mostly accounted for by the fluctuations of the moisture flux divergence (and chiefly by its velocity divergence component). An extremely high negative correlation (in the 24-hour moisture budget) was found between the horizontal moisture advection and the time change of precipitable water. A high correlation was also found between rainfall and the vertically integrated vertical moist advection. Rainfall and total precipitable water were also positively correlated. In addition there is a low, but significant positive correlation between precipitation and horizontal advection and a negative correlation between precipitation and the time change of precipitable water. Precipitation was found to be correlated with evaporation on a monthly time scale (but not on shorter time scales) indicating a possible feedback between the two variables on that time scale. Power spectrum analysis of the 12-hour moisture budget components revealed peaks in three major frequency bands, located at 2-4, 5-6, and 8-12 days. Several pairs of budget terms exhibited a high coherence squared with a nearly constant phase difference within a broad frequency band in all four summers. The most outstanding results were that maximum precipitation follows the maximum horizontal convergence term by 15 degrees and the moist advection maximum by 120 degrees. Another finding was that the time change of precipitable water and moist advection have a very high coherence squared, and are in-phase over a wide frequency band. Finally, the maximum in precipitation follows maximum precipitable water by about 20 degrees and the maximum time change of precipitable water by 110 degrees.

The goal of the ARM/GCIP measurements of soil moisture and temperature profiles (SWATS) at the ARM Southern Great Plains Site effort is to deploy a rugged, automated, and affordable soil water sensor at the surface flux stations of the ARM Southern Great Plains Site. This was also done in an attempt to meet the soil water data needs of ARM and GCIP investigators. Various factors required that a different approach from traditional efforts within the soil science or hydrologic communities be devised, requiring choosing the best currently available sensor given logistical and fiscal constraints, implementing a plan, and seeing if it worked. The implementation of this network was completed during FY 1997, while data analysis, including a thorough assessment of data quality, began during FY 1998.

Expansion and Analysis of the Comprehensive Pacific Rainfall Data Base

The objectives of the first year of this project were to digitize, in computer format, Taylor’s rainfall atlas of Pacific rainfall records going back to the 1800s, and to conduct a quality assurance exercise to weed out erroneous records. This work was completed, and a manuscript for the Bulletin of the American Meteorological Society is in preparation.

Water Vapor Measurements and Field Related Studies

CIMMS scientists installed chilled mirror dewpoint sensors at the ARM Southern Great Plains Central Facility to provide NIST traceable moisture measurement capabilities. These sensors will act as calibration standards for other moisture measurements made at the Central Facility and should provide a high level of confidence and accuracy.

These chilled mirror dewpoint sensors, installed at the surface and on a 60-m tower at the ARM Central Facility for the second Water Vapor intensive observational period, provided surface and low-level measurements necessary to complete the Raman lidar moisture profile (the lidar usually has a lowest range gate of approximately 60-m). This deployment made it possible to compare precipitable water vapor measurements made with microwave radiometers and a GPS system with those obtained from the Raman lidar. Although the shape of the Raman lidar moisture profile is correct, the absolute calibration may drift. Therefore, the 60-m tower observations were used to scale the lidar profile so that the two agree at the 60-m level. Great care was taken to ensure that the 60-m tower observations were correct; thus, the tower observations were treated as correct and the lidar profile was adjusted to agree with it. Comparisons at the Central Facility showed that microwave radiometer and Raman lidar measurements were extremely stable throughout the intensive period with respect to one another (slope = 1.01 + 0.01 with a 3 sigma standard deviation), but that there was a pure bias of ~0.18 cm between the two. The comparison to the GPS values with processing provided by the Scripps Institute of Oceanography showed a fractional difference of ~5% and only a small offset. Comparisons to ERL processing of the GPS data showed similar results, but with precipitable water vapor values that were only ~2% drier than the tower values.

Similar chilled mirror systems were also developed to support field projects requiring highly accurate, NIST traceable measurements of dewpoint. The University of Wisconsin used one system on Andros Island in 1998 as part of CAMEX-III.

Plans – FY 99

‘El Moubarak’ and Moroccan Precipitation

During FY 1998, the daily precipitation dataset will be extended backward in time. A major objective of this phase of the project will be to enhance our understanding of other meteorological processes affecting Moroccan precipitation. To that end, separate study periods will be examined in which the North Atlantic Oscillation had large or little control over Moroccan precipitation patterns. Plans for this year also include publishing at least two journal articles on past research from this project.

Predictability of East African Rains

Studies into the predictability of the East African October-December rains will continue. The findings presented at the September 1998 Kenyan Meteorological Society Workshop will be used to form hypotheses for correlations between East African seasonal rainfall totals and sea surface temperatures in both the Indo-Pacific and tropical Atlantic regions.

Investigations of Climate Variations and Trends Associated with Tropical Variability

The present funding for this project will end in December 1998. Proposals are being submitted to extend this work and to prepare journal articles.

Application of WSR-88D Data Toward Developing Four-Dimensional Wind and Cloud Fields

WSR-88D radar data from the ARM 1997 Summer SCM intensive observational period were recently obtained from NCDC. A large portion of the data received includes several radars within the ARM southern Great Plains domain. Processing of the radar data from Vance AFB (closest radar to the ARM Central Facility) has recently started using WATADS 10.0 software. Images of the reflectivity field will be saved as images (approximately 1 every 6 minutes) for the duration of the intensive period and will be collectively stored in 2- hour MPEG images for ARM Science Team use. These images will be used to help define periods of interest in which to test radar related algorithms. ADAS will be run in real time – it is an analysis package that incorporates radar and satellite data and produces 3-D cloud-related fields. Data are being retrieved from ARM to run ADAS for the intensive period under study. Much work yet remains to be able to feed the ARM data sets into ADAS. Once this is completed, analyses will be run on test cases during the intensive period. As part of this work, refinement of ADAS algorithms is expected. U.S. Department of Energy efforts using WSR-88D and millimeter cloud radar data will help this research. ADAS will likely initially be run in its current configuration, and then will likely be modified based on DOE research.

Expansion and Analysis of the Comprehensive Pacific Rainfall Data Base

Plans include for the analysis of Taylor’s digitized and quality-assured rainfall atlas. Work will include looking for trends and investigating the hypothesis that the El Niño phenomenon has been increasing in intensity. The data are will be placed on a web site for distribution to the research community. Specific objectives include exploratory data and trend analyses, with the trend analysis to assess the validity of previous work. Also, an unsupervised Neural Network analysis will be applied to the data to identify homogeneous regions of specific rainfall characteristics or regimes.

West African Monsoon

Diagnosis of West African monsoon variability will focus on the 1951-90 downtrend of monsoon rainfall in the West African Sahel, including some aspects of the interannual variability within that multidecadal trend. The variability of the distinctive West African tropospheric structure (low-level southwesterly monsoon flow, mid-tropospheric easterly wind maximum, upper tropospheric Tropical Easterly Jet) that supports the rain-producing monsoon disturbances will be investigated for the 1951-90 period using unique sets of individual rawinsonde observations, and for other decadal-scale periods using the results of GCM simulations that have already been performed (e.g., Atmospheric Modeling Intercomparison Project runs for 1979-88; decadal 1 x CO2 control runs). These findings will be related to the resulting smaller-scale monsoon disturbance characteristics (as revealed by daily rainfall data for 1951-90), the underlying larger-scale African-Atlantic basin atmospheric-oceanic-land surface conditions (e.g., meridional SST and tropospheric temperature gradients; tropical and extratropical Atlantic atmospheric circulation patterns, including the North Atlantic Oscillation, NAO; African rainfall anomaly patterns; Sahelian vegetation status) and the more global state of the climate system (e.g., tropical Atlantic versus tropical Pacific SST anomaly patterns).

North American Precipitation Variability

Diagnostic and predictability studies of precipitation variations in central and eastern North America will continue to focus on three major goals — the documentation of all aspects of the intraseasonal and interannual variability of the precipitation through the advanced statistical analysis of a fine resolution set of daily precipitation data for 1949-present; the quantification of the relative importance for precipitation amount of a range of tropospheric processes that occur over the study region, especially during growing season months; and the relation of the distribution, timing, and amount of precipitation to larger climate system controls, such as tropical Pacific sea surface temperature anomaly patterns. During the coming year, this effort will begin to “nest” a weather-system-resolving regional model within a global climate model, in an attempt to obtain maximum physical insight into the crucial links in the teleconnection chain.

Water Vapor

A temperature and relative humidity calibration chamber will be purchased by the ARM Program to provide a NIST traceable calibration facility at the Southern Great Plains Central Facility. This system will improve the temperature and moisture measurements made within the ARM area by improving the calibration schedule of the temperature and moisture sensors.

The temperature and relative humidity calibration chamber will be used to investigate the absolute accuracy of the Vaisala radiosondes used by ARM. There have been considerable lot-to-lot calibration differences in the sondes used at the Central Facility, and this work is designed to examine and perhaps eliminate these errors. NCAR has examined the calibration errors of the Vaisala radiosondes and has discovered a radiosonde “contamination” issue; that is, a relative humidity error occurs because the sondes are stored in foil bags. NCAR has developed a potential correction for this problem and Vaisala will be publishing information regarding it soon. At that time, the sonde data can be reprocessed and a correction to the relative humidity made. This problem is worst at high relative humidity and smallest at low relative humidity. Thus, even though ARM has performed sonde checks in a desiccant chamber, these checks would not have revealed the relative humidity bias because the correction is very small near 0% relative humidity. Overall, the calibration of sondes prior to launch at the Central Facility should provide important findings for the ARM Program.

Investigation of Temporal and Spatial Variations of Broadband Surface Albedo across the ARM Southern Great Plains Area

Knowledge of the broadband surface albedo, defined as the ratio of the hemispherical reflected solar flux to the hemispherical incoming solar flux, is important in modeling the surface energy budget. It is important because the difference 1-albedo is the proportion of the incoming solar flux available for sensible and latent heating at the land-atmosphere interface. For example, using a value of albedo of 0.20 when it should be 0.25 will result in a 7% overestimate of available solar flux. Clear sky albedo varies with time of day, day of year, vegetation type and health, soil type and moisture, and presence or absence of early morning dew or frost. With this background in mind, the first goal of the research is to use ARM data from clear days to investigate the space and time variability of albedo. The primary interests here are the seasonal variation of the 200-km east-to-west and north-to-south gradients of albedo and their causes. There is a 50% increase in annual precipitation from west to east across the SGP area. Since there is a significant change in vegetation and soils across the region in response to precipitation, it will be important to see if there is a corresponding response in albedo. The second goal of the research is to develop a climatology of surface albedo across the ARM SGP for all days, not just clear days. The results from the first goal should provide physical insight into why albedo varies as it does for both cloudy and clear days. The climatology would include both temporal and spatial analyses. The third goal will be to verify satellite radiance algorithms for estimating surface broadband albedo on the different spatial and temporal scales. The results of the research should be useful to atmospheric and land surface modelers to help in properly accounting for energy consumption at the land-atmosphere interface whether the models are applied to the instrumented ARM SGP area or the vast non-instrumented remainder of the Great Plains. The observed albedo can be used for both model verification and for developing new albedo algorithms in atmospheric and land surface models.

Soil Moisture from the Oklahoma Mesonet

Soil moisture has been shown to be a very important land surface characteristic that can influence mesoscale atmospheric processes. The objective of this work is to generate a reliable, self-consistent point-based soil moisture data set over the state of Oklahoma and disseminate it to the GCIP research community as soon as possible. This includes calibrating and validating soil moisture estimates from the 229-L heat dissipation sensors installed at the 112 Oklahoma Mesonet sites, and the phased release of meteorological and soil moisture data from the Oklahoma Mesonet to the GCIP research community. These data sets will be used for validation and assimilation by both research and operational modeling groups. The calibration of the sensors to soil water potential values can be done in situ and involves examination of the data streams to find minimum and maximum values over a range of moisture conditions. The major challenge we face is in accurately converting these soil water potential values to values of volumetric soil water content for sensors at all 4 depths and all 112 Mesonet sites. In the first year of this project, we will be evaluating several techniques for developing soil water retention curves that can be used for this conversion to volumetric water content. We will also begin to evaluate the spatial variability in soil moisture, including its relationship to precipitation, soil and vegetation variability.


Progress – FY 98

El Nino and its Impacts on Society

In cooperation with the Institute for Business and Home Safety (IBHS), a “white paper” on the subject of El Niño and its possible effects on the property insurance industry was prepared. The paper addressed long-term relationships with El Niño and the occurrence of floods, wildfires, hurricanes, severe weather, and nor’Easters. For each weather phenomenon, the known relationships with El Niño were examined, examples of previous insurance losses were listed, confidence levels in the year-to-year occurrence of the El Niño-related anomalies were discussed, and current predictions were outlined. A one-day workshop on the same topic was held in Washington, D.C., on November 3, 1997, for approximately 80 top executives from various components of the insurance industry. Presentations were made by Dr. Peter J. Lamb (CIMMS), Dr. Michael B. Richman (University of Oklahoma), Dr. Thomas R. Karl (NCDC), Dr. James D. Laver (CPC), and Dr. Roger S. Pulwarty (CIRES).

Climate Prediction, Information, and Policy Response: A Retrospective Assessment of Drought Management in Oklahoma

Work began on documenting prior drought and forecast information through conventional organizational and institutional channels. A regular schedule for team meetings was set, scheduling of interview trips was started, and information collection strategies were developed. A bibliographic record, including a comprehensive newspaper collection and news video coverage, was developed and materials continue to be collected. By the end of June 1998, 29 hour-long, sector-specific interviews had been conducted by the research team to profile information use and drought response. Information was gathered through telephone and personal interviews with members of the Oklahoma Governor’s Drought Management Task Force. A daylong meeting was convened by the Oklahoma Department of Civil Emergency Management, and a detailed set of information was acquired. A statewide economic impact assessment of the drought was given to the team by an agricultural economics professor from Oklahoma State University. The team began identifying environmental, economic, and social impacts associated with sector-specific impacts to estimate the value of forecast information. The Oklahoma Mesonet data base was used to create a “virtual” drought scenario for the state, and is being used to identify more accurately the geographic locations of the most severe impacts over the time period of the 1995-96 drought event. In addition, the drought team developed a wheat-growing simulation model based on the drought year to help determine the risks/benefit options available to wheat farmers in the most severely impacted areas. The modeling exercise was developed with the use of the CERES wheat model. The Oklahoma Climatological Survey research staff completed the task of characterizing the distribution of temperature and precipitation variables associated with long-range forecasts normalized for Oklahoma. The research team was represented at two drought-related research conferences. The first one was the 1997 BLM-sponsored drought conference, “Planning for the Next Drought,” in Albuquerque, New Mexico. The conference was organized by the National Drought Mitigation Center, and convened drought experts from southwestern states, including Oklahoma and Texas. A member of the research team also attended the 10th Conference on Applied Climatology of the American Meteorological Society, October 20-23, 1997, in Reno, Nevada.

Freezing Temperatures, Pipe Bursting, and Insurance Claims

Research was completed on an investigation of the occurrence of freezing temperatures in the southeastern U.S. and their relationship with insurance claims and losses due to pipe bursting. The arrival of a proprietary set of insurance claim data in May 1996 initiated the study of claim occurrence in relation to earlier ongoing analyses of the freezing temperature variability in the Southeast. This work initially culminated in a Master’s Thesis by Brian Skinner. It included results of each of the following: 1) analysis of the climatology of freeze occurrence in the southeastern United States — mean occurrence, standard deviations, linear and parabolic trends in occurrence, and analysis of spatial variability; 2) analysis of the spatial and temporal variability of pipe-bursting-related insurance claims; 3) daily analysis and comparison of two catastrophic pipe-freezing events; 4) derivation of alert indicator temperature thresholds for pipe-freezing for distinct multi-county regions; and 5) application of the indicator temperature thresholds in economic analyses, a study of the climatology of severe-freezing temperatures, and the development of a loss-prediction method based on predicted monthly temperatures. Beginning in May 1997, the above research results were formulated into a set of reports to the Institute for Business and Home Safety (IBHS). New data were received in May 1997, which were added to the existing results, and a revision process was initiated based on comments from IBHS and insurance-industry representatives. A final report to the IBHS on the investigation of freezing temperatures and their relationship with insurance claims and losses due to pipe bursting was completed and submitted in February 1998.

Agricultural Production and Economic Decision Models

A whole-farm level decision model was used to examine the impact of the type of decisions producers make on the value of seasonal climate forecasts for east central Texas. Results suggested precipitation forecasts directed toward crop mix and applied nitrogen level decisions would have the largest economic value. Further, the results showed that the economic value of climate forecasts could not be obtained by examining only a small set of decision types. Rather, all decision types must be modeled to correctly value seasonal precipitation forecasts. This occurs because, in response to seasonal climate forecasts, changes in one type of decision may override the need to change other decision types. Finally, forecasts of precipitation during the crop tasseling and grain filling stages may be more valuable than precipitation forecasts for earlier crop growth periods.

Economic decision models incorporating biophysical simulation models were used to examine the impact of the use of Southern Oscillation (SO) information on sorghum production in Texas. Production for 18 sites was aggregated to examine the impact of the use of SO information on the aggregate supply curve and other economic variables. Two information scenarios were examined. The first scenario assumed producers do not use SO information in making their production decisions. This scenario was contrasted to a scenario in which producers use information concerning the SO in making their production decisions. For all expected prices, the use of SO information increased producers’ net returns. The expected Texas aggregate sorghum supply curve using SO information shifted both left and right of the “without information” supply curve depending on the price. Changes in nitrogen use based on the SO information were a major factor causing the shift in the supply curves. Further, the use of SO information decreased aggregate expected costs of production. Changes associated with the use of SO information can be summarized as follows — the use of SO information provides producers a method to use inputs more efficiently. This more efficient use has implications for both the environment and for the agricultural sector.

Time series models were used to examine the impact of Southern Oscillation (SO) extreme events in estimating and forecasting Texas sorghum and winter wheat yields. Results were both crop and period specific. Including SO events in forecasting yields decreases the forecast mean square error for winter wheat but had no significant impact on sorghum forecasts. Further, it was shown that a significant correlation between SO events and yields does not necessarily translate into better yield forecasts.

Plans – FY 99

International Center for Natural Hazard and Disaster Research (ICNHDR)

Natural phenomena such as floods and droughts become hazards when society is impacted. Droughts and floods are normal recurring natural phenomena with increasing socioeconomic impacts in developing as well as developed countries. Natural hazard and disaster research is concerned with the prediction and characterization; risk and vulnerability; mitigation; and response and recovery in regional, national, international, and global contexts. The subject matter areas of the new International Center for Natural Hazard and Disaster Research (ICNHDR) are those hazards posed by water; either too much (floods) or too little (droughts). The methodological bases of Center research will draw from distributed parameter simulation models, coupling of hydrologic and atmospheric models, adjoint optimization techniques, and spatial information management. Engineering, meteorological, and climatological disciplines will be the primary intellectual and professional disciplinary focus. Engineering is primarily concerned with infrastructure and technology designed for the benefit of society. Physical science seeks to characterize and understand natural phenomena. ICNHDR, through research, will promote the development of technology and better-engineered systems for reduction of impacts to society posed by natural hazards. Technology transfer will occur through research and development of software, system integration, and instrumentation for flood warning systems, coupled model system software for predicting hydrologic impacts of severe weather. Education and training will be provided on natural hazards and disasters for municipalities, agriculture, consulting engineers, and weather related industries.

Flooding is responsible for greater annual costs than any other natural disaster. As density and location of population within flood prone areas increases, so do flood damages. Flood warning systems are often the only means communities have to reduce vulnerability. Fundamentals of catchment hydrology and hydraulics, as well as, numerical methods, computing, and information management systems are essential to the development of simulation models for predicting the hydrologic response of urban and rural areas. New sensor systems such as mesonets, radar and satellite offer improved characterization of flood hazard potential in real time and post analysis. Concomitant hazards associated with heavy rainfall, severe storms and flooding include wind damage, erosion, sedimentation, landslides, and storm surge. Climatological flooding induced by consistently higher levels of rainfall produce waterlogged soils, and changes in the hydrologic response over large regions. Understanding the short term, local-scale flooding, as well as, long-term regional scale flooding will be the focus of the Center’s effort.

Short to long term climatological drought is identified as a period of insufficient water resources initiated by reduced precipitation. Shifts in local or regional precipitation patterns have important impacts on the hydrologic response. Key areas of Center research will include understanding the initiation, perpetuation, and termination of drought; development of predictive models, warning and response systems; and the impacts on society. Hydrologic response to regional and global climate change coupled with changes in population location and density will continue to aggravate national and international vulnerability in terms of water resources and food security as witnessed by the recent drought in various parts of the world such as sub-Saharan Africa.

Climate Prediction, Information, and Policy Response: A Retrospective Assessment of Drought Management in Oklahoma

An unexpected opportunity arose for the research team to undertake a comparative analysis of drought-based risk perception and information use. With the advent of the 1998 drought in Oklahoma, the team has decided to re-orient its approach to include the recent and continuing drought-related events, particularly in the areas of agriculture and wildfire suppression. In the second year of the project, the team will develop and administer a survey questionnaire. The significant difference that has occurred in the last year is the growing importance of the Oklahoma Mesonet in weather-related information dissemination and documentation. This new aspect of the 1998 drought will be a key research activity for the team. The team will develop additional historic fire data and wheat and cattle production data. The State Fire Marshall’s office has agreed to allow the team to define fire occurrences by geographic location, so that a complete statewide data set can be developed for analysis with the Mesonet-based drought model. The completed modeling and mapping work will be modified to fit the specific questions designed for the survey instrument.

Modern Meteorology and the Insurance Industry

In the next years, a significant enhancement of a carefully structured dialogue between the meteorological and the insurance communities will take place in partnership with the IBHS. This dialogue will focus on how the capabilities of modern meteorology can be used to reduce the toll on life and property resulting from severe weather and climate change, and enhance the competitive positions of proactive primary insurance companies. It will explore how the insurance industry can better use weather and climate information in strategic planning, in anticipating and mitigating losses of insured property, and in operational decisions responding to weather-related catastrophes.

Agricultural Production and Seasonal Climate Prediction Models

The long-term competitiveness of United States agriculture centers on its adoption of technological advances. Recent developments in the science of meteorology are providing the means to more accurately predict seasonal climatic conditions. Improved forecasts will alter producers’ selection of inputs and expectations of output price. Such changes would affect cropping patterns and the use of production inputs that subsequently alter regional and national supply curves. These changes will impact the competitiveness of the United States in international markets. Currently, Australia has improved seasonal climate forecasts that are being made available to the private sector decision-makers. To our knowledge, no study has focused on how these improved climate forecasts affect international competitiveness. The main objective of this proposed study is to address this void by evaluating the role of enhanced climate forecasting ability on international competitiveness in the world grain market. This work is being performed in collaboration with agricultural economists at Texas A & M University.


Progress – FY 98

Open Systems Radar Product Generator (ORPG)

NSSL was tasked several years ago to transition the WSR-88D Radar Product Generator (RPG) to an open systems environment, or Open Systems RPG (ORPG). A total of five Incremental Software Builds were planned for the completion of the ORPG. Two of these builds were completed during the progress period. Reviews and demonstrations of the ORPG were conducted in July 1997 (Incremental Build 2) and January 1998 (Incremental Build 3). Both reviews were well attended by WSR-88D experts and stakeholders. Consensus continued to affirm significant progress continues to be made with the project and that the ORPG will serve the growing operational demands of “tri-agency” users (Department of Commerce, Department of Defense, and Department of Transportation) quite well. Official software and documentation releases of each incremental build were delivered to the OSF for testing and review. Engineering software drops of near-Build 4 functionality were provided to the OSF through June 1998 to support ORPG hardware selection. ORPG team members have been accessing data from the NSSL/NOAA WSR-88D for ORPG development and testing purposes. The ORPG can currently perform most of the control and monitoring functions of the legacy system. The Product Distribution function is considered to be one of the more complicated components of the legacy system. Considerable effort has gone into simplifying its design to increase the level of maintainability as well as enhance the flexibility and extensibility of its function. To date, all legacy algorithms and products (51 in all) have been ported and integrated into the ORPG. These applications represent the only software being ported from the legacy system. The remainder of the ORPG, including supporting software infrastructure and system applications, represents entirely new development. A very critical enhancement to the ORPG is the development of a new Human/Computer Interface (HCI) to replace the Unit Control Position (UCP). The existing UCP is an alphanumeric, nested menu interface that is not intuitive and requires significant training. The new HCI takes advantage of an intuitive, graphical interface to display system status and control. To date, 27 of a prescribed 40 HCI screens have been developed.

A Masters thesis (Mr. Michael Rausch – OSF Applications Branch) in industrial engineering was completed in the area of information systems design. It is titled “Design and Evaluation of a Graphical User Interface for the WSR-88D Control Software.” The primary task was to oversee development of a graphical user interface, HCI, as a replacement for the WSR-88D UCP, as described above. Mr. Rausch assembled a team of OSF staff members and an industrial engineering undergraduate student to develop and employ quantitative testing methodology. Tasks included gathering input from users of the interface, designing prototypes, testing the prototypes using OU meteorology student volunteers and OSF staffers, displaying the prototypes on the Web, and communicating results to the project leader and head programmer.

Open Systems Principal User Processor (OPUP)

NSSL was tasked to develop an Open Systems Principal User Processor (OPUP) for Department of Defense users of the WSR-88D system. The Principal User Processor is a dedicated weather radar display system that is currently deployed as part of the WSR-88D equipment. NSSL began development activities for the OPUP system, including a review of requirements, a review of available ORPG software infrastructure and support services, and development of prototype applications. Prototype applications were developed for communicating with the WSR-88D RPG, product storage, graphical product display and control, and configuration functions.

Statistical Analysis of WSR-88D Level II Data

Another Masters thesis student in industrial engineering (Mr. Chidambaram Jeyabalan – OSF Applications Branch) completed a thesis in the area of WSR-88D data analysis. It is titled “Statistical Analysis of the Base Data from the WSR-88D Weather Radar and Description of Mesocyclone Features Using Neural Networks”. At least two papers have been published on his work. Level II data were explored for their structure. Using sophisticated multivariate statistical techniques, patterns were identified the in raw data that are known to contain mesocyclonic circulations and tornadoes. The current WSR-88D mesocyclone and tornado algorithms use only velocity information, while the new techniques developed used all three moments: velocity, reflectivity, and spectrum width. This work proved promising and may lead to a new class of statistically based algorithms in much the same way that statistical thermodynamics advanced physical chemistry.

Advanced Weather Interactive Processing System (AWIPS)

The NWS, as part of their modernization activities, is deploying the Advanced Weather Interactive Processing System (AWIPS) to all forecast offices. This new system will allow an integrated sensor approach to detection and warning for severe thunderstorms and tornadoes, including radar, satellite, lightning and other sensor data. These data will be blended to help arrive at warning decisions. This represents a shift in NWS use of radar in warnings (which had been used more as a stand-alone sensor) and in the type of training previously provided by the OSF (the PUP). CIMMS, working with OSF, is involved in new efforts with AWIPS in two ways: first, development of AWIPS capability at OSF/OTB, and second, development of training associated with new integrated sensor applications. For the first objective, AWIPS software is being ported to OSF/OTB UNIX workstations and optimized to run on a network of applications and training development systems. Multiple data types and complete data sets are being prepared to run on the UNIX network in the same manner in which the data will be processed in field-office AWIPS units. For the second objective, stand-alone radar signatures and algorithm outputs are being integrated into multi-sensor applications (particularly, joint use of radar and satellite information) and tested on storm and tornado cases from 1997 and 1998. Important activities have included:

  • Developing a better understanding of TVS evolution
  • Developing a better understanding of how WSR-88D radars see tornadogenesis
  • Using radar and satellite to reveal tornadic storm updraft evolution
  • Using radar and satellite to better estimate thunderstorm rainfall
  • Developing joint radar and satellite methods for determination thunderstorm equilibrium level.
  • Analyzing the Jarrell, TX tornado of May 27, 1997
  • Analyzing the Hall-White Counties, Georgia, tornado of March 20, 1998
  • Working with NSSL, as part of SCAN, to design and test next-generation AWIPS software

Polarimetric Radar Developments

Regular precipitation observations are now being collected with the NSSL’s dual-polarization Cimarron radar. Thirty-two storms have been observed during the period from 1 July 1997 to 30 June 1998. In addition, a new version of the polarimetric algorithm for rainfall estimation based on joint use of specific differential phase KDP and differential reflectivity ZDR has been developed and tested for 15 storms. A new algorithm for areal rainfall estimates based on a contour integral of a total differential phase also has been developed and tested on 14 Oklahoma storms. Also underway is a comparison of radar polarimetric data and disdrometer precipitation data for 25 precipitation events in central Oklahoma.

In collaboration with scientists from the National Center for Atmospheric Research (NCAR), work was completed on a real-time polarimetric radar algorithm to identify and classify hydrometeor types. The algorithm uses fuzzy logic to weight and combine polarimetric radar measurements of reflectivity (Z), differential reflectivity (Zdr), specific differential phase (Kdp), cross correlation coefficient (rhv (0)), and linear depolarization ratio (Ldr). The algorithm also identifies radar signatures associated with anomalous propagation, birds, and insects. NSSL and CIMMS recently scientists participated in the TExas/FLorida Underflight (TEFLUN-B) Experiment near Melbourne, Florida. Though some fine-tuning of the algorithm is needed, an initial evaluation of its performance on the NCAR S-POL radar during that experiment indicates that the algorithm effectively determines hydrometeor types.

Long-term statistics of polarimetric observations in winter storms in central Oklahoma were examined to develop the procedure for rain/snow discrimination as a part of a general classification algorithm. The zones of negative specific differential phase KDP near the tops of clouds were also identified and examined as possible indicators of vertically aligned crystals due to the presence of strong vertical electric fields.

A comparative study of different schemes for implementation of polarization diversity on the WSR-88D radar was performed. The contribution was made to the NSSL’s report “Polarimetric Upgrades to Improve Rainfall Measurements” (prepared by D. Doviak and D. Zrnic – April 1998).

Mitigation of Velocity and Range Ambiguities

CIMMS scientists continued investigation of schemes to mitigate range and velocity ambiguities. Four different phase coding methods for the recovery of spectral moments of overlaid weather signals have been evaluated. Three of these, (a) the random phase coding, (b) the p /4 phase coding, and (c) the p /2 phase coding have previously been published by CIMMS and NSSL authors. The fourth is a set of codes proposed by Sachidananda and Zrnic for which henceforth we use the abbreviation SZ (n/M), where M is the code length and n is an integer which defines the spectral properties of the code. To evaluate the performance of the phase coding methods, simulated weather signals were used in a versatile program that generates any desired spectral parameters and coding scheme. Simulations include practical effects of noise, data window, phase uncertainty, quantization, code synchronization, and presence of ground clutter. The simulation studies indicate that the proposed SZ (n/M) phase coding scheme performs better than any of the other coding schemes, and is the best candidate for implementation on the WSR-88D. The uniform PRT used in the phase coding scheme allows effective filtering of the ground clutter. A decoding algorithm has been developed to estimate all the spectral parameters of two overlaid weather signals. The algorithm uses processing in the spectral domain and, hence, needs a powerful radar signal processor.

Development of a Multiple Pulse Repetition Frequency Dealiasing Algorithm (MPDA)

Work continued toward a final version of the Multi-PRF Dealiasing Algorithm (MPDA). This algorithm involves using new WSR-88D scanning strategies and processing to mitigate both range folded velocity data and velocity dealiasing errors. WSR-88D sweeps are collected at different Nyquist velocities while maintaining a constant elevation angle. The sweeps are then merged and final velocity values calculated. This has proven to provide less range folding and better velocity products than utilizing only single sweep data. During the 1997-98 time period, the prototype algorithm was continuously improved upon. Data were collected using scanning strategies developed from previous MPDA analysis.

Rapid Update of Output from the NSSL WSR-88D Algorithms

One drawback of the current WSR-88D system is that algorithm processing is not done until after the completion of each volume scan of data. Thus, vital algorithm information such as circulation position updates and detections that may exist at the lower radar elevation angles are not available until the radar finishes scanning through the higher elevation angles. This lag time can be several minutes. NSSL/CIMMS scientists, through funding from the FAA, have developed software that updates algorithm output after every sweep. Thus, new severe weather detections can now be immediately relayed to the user.

The Storm Cell Identification and Tracking (SCIT) algorithm has been modified to produce intermediate output before the end of the volume scan. Expanding on the “rapid update” feature, storm cell information can now be displayed as each elevation angle is completed. After data from the first elevation angle are received, two-dimensional storm cell components at the lowest elevation are matched with cells from the previous volume scan, and their positions are updated. The “new” 2-D components inherit the attributes of the associated “old” storm cell until a new 3-D cell is built and new attributes are computed. Above the first elevation angle, two-dimensional storm cell components at adjacent elevations angles are vertically associated and combined into new 3-D features, according to standard algorithm criteria, as each elevation scan is completed. When a storm cell ceases to have new 2-D components added, it is considered “topped.” Topped 3-D cells are matched with 3-D cells from the previous volume scan and feature attributes are calculated. Associated with past 3-D cells allows for tracking information (previous and forecasted locations) as well as a feature identifier to be assigned. Topped 3-D features are then identified on the radar display as each elevation angle is completed.

Correlation Tracker/Scale Separation/Lightning Warning Algorithm

Software packages that perform correlation tracking and scale separation were obtained from MIT’s Lincoln Laboratory and incorporated into a Lightning Warning Algorithm package for WeatherData, Inc., in Wichita, Kansas. The package inputs gridded radar data into the Scale Separation and Correlation Tracker algorithms, which separate large-scale features (such as squall lines) from individual storm cells and calculate motion vectors based on the degree of correlation in the data between successive images. The Lightning Warning Algorithm uses output from these two algorithms and National Lightning Detection Network data to predict the probability of lightning in 10, 20, and 30 minutes. The algorithm package is nearing completion, and the testing phase will begin soon. The feasibility of incorporating Scale Separation and Correlation Tracker algorithms into the WDSS via a comparison study with SCIT output needs to be determined.

Damaging Downburst Prediction and Detection Algorithm (DDPDA)

The Damaging Downburst Prediction and Detection Algorithm (DDPDA) was completely rewritten in the C programming language. Additional methods and programs were developed to help better assess the performance of the algorithm and to facilitate easier data collection. The primary events examined were high reflectivity cells that were observed either close to a radar, where the strength of the outflow can be accurately measured, over a wind sensor, or over a highly populated area (where damaging winds are most likely to be reported). The DDPDA was evaluated using a data set of 226 cells located over populated areas, 30 of which were severe. In its current state, the algorithm yielded 20 hits, 10 missed events, 21 false alarms, and 175 correct non-forecasts of non-events. The probability of detection was 0.617, while the false alarm ratio was 0.512 and the critical success index was 0.392. On average, the algorithm gave an 8.5-minute lead-time for events in this data set. CIMMS staff at NSSL facilitated agreements between the TDWR Program Support Facility (PSF) and the Low Level Windshear Alert System (LLWAS) Program Office to acquire TDWR and LLWAS microburst/downburst algorithm output (in and around 10 major U.S. airports), data that are necessary for verification. Additional agreements have been initiated for obtaining various mesonet data for detecting possible strong microburst/downburst events. Linear multivariate discriminate analysis was used on the available data to determine which WSR-88D derived parameters are useful for detecting damaging events.

Airborne Doppler Data Analysis and System Development

From May 15-June 15, 1998, during MEaPRS, NSSL staff operated the radar onboard the NOAA P-3 aircraft. Just before the MEaPRS experiment began, the NOAA Aircraft Operations Center upgraded the radar system onboard the P-3 to enable it to operate in a dual PRF mode. The first experimental mission of the MEaPRS experiment was performed to test the new radar system. Analysis of the data from the test flight has shown that the new system now affords a Nyquist velocity of about 51 meters per second as compared to the previous Nyquist velocity of around 13 meters per second. This radar upgrade will significantly reduce the amount of time required to edit and analyze airborne Doppler radar data. The new system was used routinely on all subsequent P-3 flight during MEaPRS.

Work has continued on the P-3 airborne Doppler radar data collected during the Fronts and Atlantic Storm Track EXperiment (FASTEX). The field portion of the experiment, which included four aircraft (one NOAA P-3 and one NOAA Gulfstream IV) and numerous surface vessels (one or more of which was NOAA ships), was conducted out of Shannon, Ireland, during January-February 1997. A total of 18 experimental missions were conducted and to date, airborne Doppler data have been analyzed from three of the missions. Work continues on these data sets, including both individual P-3 data sets and a combination of NOAA P-3 and NCAR ELDORA data sets.

From January 15-March 30, 1998, the Caljet experiment was conducted from a base in Monterey, California. It also included NSSL staff operating the radar onboard the NOAA P-3 aircraft. The experiment was initially conducted to study land falling low-level jets that strike the western coast of the United States. The experiment was scheduled to last six weeks, from January 15-February 28. As the scheduled ending time approached, a request was made to extend the project for another 65 hours of NOAA P-3 flight time. The extension was requested because the real-time dropwindsonde data provided by the aircraft had enhanced the forecast models significantly during the then ongoing El Niño event.

WATADS Development

The performance of WATADS was monitored and improved during the past year by a member of the OSF Applications Branch. WATADS is a software package designed to emulate the WSR-88D radar and is used for analysis of Level II data.

SWAT-V Severe Weather Algorithms for the WSR-88D

During 1998, the Severe Weather Warning Applications and Technology Transfer – Vortex focus group (SWAT-V) was formed at NSSL. Its mission is to develop thunderstorm-scale vortex detection applications. These applications are primarily for the WSR-88D radar, but they also integrate data from other sources (mesoscale models, surface data, etc.) and help transfer technology and knowledge from NSSL to on-line NWS and FAA forecasters to better enhance their severe weather warning and tornado warning decision making. Work on the following algorithms took place during the past year:

  • Mesocyclone Detection Algorithm (MDA): The NSSL MDA was modified to output information in a “rapid update” format, at the end of each elevation scan, rather than at the end of a completed volume scan (in which the data might be up to 5 minutes old). This allows NWS and FAA forecasters to obtain detections from the MDA as soon as detection criteria are met.
  • Tornado Detection Algorithm (TDA): The NSSL TDA was also modified to output information in a “rapid update” format. The NSSL assisted the OSF Applications Branch with the implementation of the TDA into the Build 10 Radar Products Generator (RPG). This included consultation regarding the TDA functionality and analysis of output parameters from the TDA to determine guidance for NWS/FAA field use. In addition, NSSL and NWS personnel collaborated in the related analysis of descending and non-descending tornadic vortex signatures. A draft of the manuscript (R. J. Trapp, lead author) was submitted to Weather and Forecasting for review.
  • Neural Networks and Statistical Analyses: The development of Bayesian-type Neural Networks (NNs), though in progress, has taken secondary place to the development of the non-Bayesian NNs. The latter NNs were given emphasis because their rate of maturity has been sufficiently rapid to warrant their “release” as part of the SSAP. Three NNs have been released – one for MDA detection, a second for TDA detection, and a third for circulations detected jointly by the MDA and TDA. Each NN has two output nodes – the posterior probability of tornadoes and severe winds, respectively. These NNs were developed using a data set containing 29 tornado days from across the U.S., with 93 MDA and TDA detection attributes and ground truth verifications. The 29-day dataset was subsequently expanded to 43 cases, and new MDA and TDA detection attributes (e.g., mesocyclone descent/ascent rate from the Trapp et al. study) were incorporated. Analysis included the association of ground truth verification data to individual algorithm detections. Also, Bounded Weak Echo Region (BWER) algorithm attributes were included in the data set (for a grand total of 119 detection attributes). This data set is currently being used as a new NN training set to eventually replace the 29-case NN described above.
  • Vortex Detection and Diagnosis Algorithm (VDDA): A pilot study was completed regarding the identification of azimuthal and radial shear regions within WSR-88D data using a least squares derivative technique. This technique was envisioned to be useful in the identification of vortices associated with tornadoes and mesocyclones, as well as identification of divergence associated with downbursts. Also, the technique should be able to differentiate false detections due to linear shear phenomenon (e.g., gust fronts) and true vortex detections. Plans include incorporating a least squares derivative technique for the identification of azimuthal and radial shear regions into a future version of the VDDA. Also, simulated WSR-88D data were created for analytical Rankine vortices with varying diameters, rotational velocities, range from the radar, and beam center-to-vortex offsets. These data have been used to test the MDA and TDA for analytical difference, and algorithm strengths and weaknesses. The least-squares derivative technique will also be tested on these simulated data. Also, special data sets of simulated mesocyclones with embedded TVSs, mesocyclones with associated rear-flank downdraft gust fronts, and pure linear gust fronts, all with varying strengths, sizes, ranges, and orientations, were developed for testing. The goal of the testing is development of a new two-dimensional vortex feature detector that detects all scales of thunderstorm-scale vortices (mesocyclone and TVSs) for the VDDA. Tests are currently underway and will be completed in 1999.
  • Tornado Warning Guidance: A conference paper was produced to disseminate the findings of an analysis of the 29-day data set designed to isolate the “best” predictors of tornadoes for the development of Tornado Warning Guidance documentation for the NWS.
  • Vortex Climatology: Two WSR-88D sites were chosen for initial climatology analyses: Pittsburgh and St. Louis. Complete data for 1997 at Pittsburgh have been collected. Processing has begun and is progressing faster than anticipated. Data were divided into two groups: data from events associated with damage (as determined using the SPC Smooth Log data set) and data from events not associated with damage. Events have been subdivided into two groups as well: algorithm output associated with TVSs and data associated with mesocyclone circulations. Statistical characterization of these began as soon as data became available.
  • Formal Issues in Forecast/Warning Verification: A mathematical analysis was performed of measures of performance that are linear in the loss-matrix. It was shown that if the loss matrix is nonsymmetrical, as is the case in most realistic situations, then there does not exist a unique performance measure that would satisfy the constraints of equitability, as put forth by Gandin and Murphy. In short, it was shown that there does not exist a unique, equitable skill score for two-category events that have unequal costs associated with a miss and a false alarm. A paper has been accepted for publication in Monthly Weather Review.

Data Assimilation

Existing Doppler wind and thermodynamic retrieval packages in the ARPS ADAS were tested with the NRL COAMPS. Currently, the codes are being upgraded with terrain coordinates using COAMPS backgrounds.

Plans – FY 99

Open Systems Radar Product Generator (ORPG)

Software development for the ORPG at NSSL will be completed during the coming year. Incremental Build 4 is scheduled for delivery to the OSF in September 1998. The final software release of the system (Incremental Build 5) is expected during the first quarter of calendar year 1999. Following the delivery of Incremental Build 5, the ORPG Development Team will work with the OSF to perform system-level testing and begin preparation of the system for fielding. Deployment of the ORPG is currently scheduled for FY 2000.

Open Systems Principal User Processor (OPUP)

Development of the OPUP system continues, with the scheduled release of the first software mini-build, which includes basic functionality. Additional functionality and refinements of earlier developments will be added in subsequent mini-builds.

Advanced Weather Interactive Processing System (AWIPS)

Development of OSF/OTB AWIPS capability will be completed and OTB distance learning courses will be transitioned to an AWIPS baseline instead of a PUP baseline. Additional integrated-sensor data sets will be developed and used with new AWIPS applications and in new training materials. During the next year, the new Open Systems RPG will be completed, and it will need to be integrated into the OTB suite of systems. Training materials for the Open Systems RPG will need to be developed. For tornado cases analyzed during the last year, on-line training modules will be prepared to instruct all NWS forecasters on how to better utilize radar and other sensor data in difficult warning situations. Utilizing new integrated-sensor applications, new training workshops will be developed on Warning Decision-Making, Pulse Storm Downburst Prediction, and Quantitative Precipitation Estimation. These workshops will be delivered at COMET in Boulder, Colorado, during the next year. In cooperation with NSSL, there will be continued participation in the development of next-generation WDSS software. As part of the SCAN Program, the software will tested at multiple NWS sites, in Oklahoma and in other parts of the country. WDSS training materials will also be developed.

Polarimetric Radar Developments

Regular precipitation observations and data collection with the Cimarron radar will continue. Work to be done includes estimation of the relative effects of the drop size distribution and shape variability on the performance of the polarimetric method for rainfall estimation. Theoretical analyses will include analysis of the data obtained from the NSSL disdrometer in previous years and the data collected from the 2-D video disdrometer in the framework of the MEaPRS experiment. The latest version of the polarimetric rainfall algorithm with new ground-truth data collected from the Agricultural Research Service (ARS) Micronetwork will also be tested.

Work will continue to refine weighting functions and improve hydrometeor discrimination algorithm performance. NSSL will take part in the analysis of the SPOL observations made by NCAR in Florida as part of TEFLUN-B. Data collection is planned again during 1999. Verification will also take place using the case of May 17, 1995, for which radar polarimetric data and in-situ T-28 aircraft measurements are available. Work on this algorithm will entail an exhaustive comparison of the algorithm performance on a large number of precipitation events contained within the NSSL Cimarron radar archive (consisting of over 200 tapes of both cold and warm season precipitation events). Work will also be conducted to develop methods to quantify hydrometeor amounts. Finally, the algorithm will be implemented on the NSSL Cimarron and prototype polarimetric WSR-88D radars.

The possible relation of negative KDP signatures measured by the Cimarron radar will be examined with lightning flashes collected by the National Lightning Detection Network.

A methodology will be proposed to test the new polarimetric scheme on the WSR-88D, and a test of the prototype of the polarimetric WSR-88D being developed at the NSSL will take place.

Development of a Multiple Pulse Repetition Frequency Dealiasing Algorithm (MPDA)

Testing and enhancement will continue with the MPDA package. Scanning strategies are planned which will maximize velocity recovery while minimizing velocity dealiasing errors. Data collection is planned throughout the period using the OSF KCRI WSR-88D. The ultimate MPDA goal for the 1998-99 time frame is to finalize the MPDA package and deliver it to the OSF for inclusion on the WSR-88D.

Rapid Update of Output from the NSSL WSR-88D Algorithms

Real-time testing and enhancements of the “rapid update” software are planned during the next year. The eventual goal of this software concept is integration within the WSR-88D system.

Other WSR-88D Radar Quality Improvements

Activities scheduled include investigation of regression filters as ground clutter filters in the context of variable PRT and investigation of parameter estimation techniques for data quality improvement.

Volume Coverage Pattern (VCP) Testing

Existing VCPs in the WSR-88D system are not meeting the needs of all users. In particular, there is a demand for volume scans with a faster total update time for use during severe weather events. The OSF has tasked the NSSL to develop and test such a volume scan. A 3.5-minute volume scan was created and tested using existing Level II data tapes that contain data collected in VCP 11 (fourteen elevation angles in five minutes). The Level II tapes were modified to decrease the total volume scan time by deleting several elevation angles such that the remaining elevation angles were equivalent to those used in VCP 21 (nine elevation angles in six minutes). Two events have been designated as test cases: Melbourne, Florida (KMLB), 25 March 1992, and Frederick, Oklahoma (KFDR), 8 May 1993. Current WSR-88D algorithms and NSSL developmental algorithms are being run on the original and modified data sets. Results from both data sets are being compared to assess the feasibility and usefulness of the proposed 3.5-minute volume scan. Other operational needs exist that can be solved by creating new or modified VCPs for the WSR-88D. The second part of this project consists of identifying those needs and determining solutions (new or modified VCPs). A “VCP Working Group”, which includes CIMMS employees, has been established to do this. This group meets every couple of weeks for discussion. This group is a combined effort between the OSF and the NSSL.

Incorporation of Terrain Information into the Storm Cell Identification and Tracking (SCIT) Algorithm

Information from a terrain database (such as ArcView or GRASS) will be incorporated into the SCIT algorithm. The database will be consistent with any terrain information being used in the existing WSR-88D system. Code will be incorporated into the SCIT algorithm to reference the terrain database when making calculations that involve height. The results of the incorporation will be height calculations AGL (from the terrain information) instead of using the assumption of a flat earth. These calculations will feed directly into the Hail Detection Algorithm (HDA). Algorithm tuning and performance evaluation before and after the addition of terrain will represent another goal. In addition, a preliminary evaluation will be made on what impact the incorporation of terrain into SCIT has on other algorithms that are dependant on SCIT.

Damaging Downburst Prediction and Detection Algorithm (DDPDA)

The Damaging Downburst Prediction and Detection Algorithm (DDPDA) will be expanded to include prediction capabilities for cells that form in an environment of low CAPE and a deep boundary layer with a steep lapse rate. Severe convective wind events that occur in this type of environment are frequently called “dry downbursts.” This effort will require the evaluation of available environmental data (such as NSSL’s Near-Storm Environment algorithm) and a change in the default SCIT algorithm parameters to enable SCIT to detect weaker cells for processing by the DDPDA. Additional work will involve the expansion of the downburst database to include data from the higher-resolution Terminal Doppler Weather Radar (TDWR) as verification for cells observed with the WSR-88D. These data will be used in the discriminant analysis process, and will possibly be used to guide more sophisticated Neural Network approaches to detection and prediction. Also, verification data will be binned within particular ranges as tests are performed to determine the variability between WSR-88D derived parameters with range and, should the data set become large enough, geographic region. Also, the value of various existing and proposed detection/prediction algorithms will be investigated.

Airborne Doppler Data Analysis and System Development

Work will continue on the analysis of FASTEX airborne Doppler data as well as data collected during the MEaPRS field program. In addition, work will continue on evaluation of the new radar system onboard the NOAA P-3. With the Mesoscale Alpine Program (MAP) scheduled to begin in September 1999, NSSL will be involved with and assisting in flight pattern planning. MAP is an international project to be staged from Innsbruck, Austria, with NSSL providing expertise in airborne Doppler data collection and airborne mission coordination.

SWAT-V Severe Weather Algorithms for the WSR-88D

  • Mesocyclone Detection Algorithm (MDA): The “rapid update” version of the algorithm will be tested at a NWS real-time site during fall 1998 or spring 1999. Also, the Near-Storm Environment (NSE) data from the RUC-II model will be incorporated into the MDA. These data will replace the conventional rawinsonde data having 12-hour, 300-km resolution with 1-hour, 40-km resolution.
  • Tornado Detection Algorithm (TDA): The “rapid update” version of the algorithm will also be tested at a NWS real-time site during fall 1998 or spring 1999. Near-Storm Environment (NSE) data from the RUC-II model will be incorporated into the TDA (see MDA above). The Build 10 TDA is planned to be installed at the WFOs during fall 1998. NSSL will act in a consulting role to continue to help develop training materials and to field any questions on the use of the new TDA in NWS warning operations.
  • Neural Networks and Statistical Analyses: The training of an NN on the 43-case 119-attribute data set will be completed. Also, another NN will be developed on a subset of the 43-case data set that has NSE data (RUC-I mesoscale model) incorporated as additional attributes.
  • Vortex Detection and Diagnosis Algorithm (VDDA): The two-dimensional feature detector will be completed. New three-dimensional (vertical) association and time association techniques will be developed. The VDDA will be tested in real-time and on canned datasets during 1999 and 2000. NN development will begin on VDDA data. Data will also include BWER algorithm output and NSE information.
  • Tornado Warning Guidance: Similar analyses will be performed on a 43-day data set to provide a more reliable assessment of the question of the “best” predictors of tornadoes.
  • Formal Issues in Verification: Analyses so far are incomplete in many respects and will therefore be continued.
  • Vortex Climatology: Data retrieval of Pittsburgh and St. Louis algorithm output will be completed. Statistical characterization of circulations in general, circulations associated with TVS algorithm output, and circulations associated with mesocyclone-related algorithm output, will be made. These statistics will be used to characterize real-world characteristics of WSR-88D algorithms and what they detect in convective weather. This can, in turn, be used to examine the quality of observational data, and determine if they can be useful for better interpreting algorithm output. These statistics can also be used to check the algorithms themselves – obviously erroneous values can be flagged and then algorithms checked as to the cause. Differences in climatologies between sites can also help tune algorithm output for best performance, using site-specific parameters.

Data Assimilation

Future work will involve development and testing of wind-thermodynamic retrieval algorithms using the COAMPS background with terrain coordinates. The spline smoother will be compared against the penalty-function smoother for possible improvements in the retrieval schemes