The End User Integration (EUI) team incorporates user needs into system design and links CASA's technology to individuals, organizations, and society at large.

The EUI team interacts with users of weather data, such as the National Weather Service, private meteorology companies, academic researchers, and emergency managers. The team studies social, policy, behavioral, and technical interface issues around the use of DCAS systems in weather impacted decision making and response.

The goals of the EUI team focus on four interrelated areas:

1. Specification of DCAS Adaptive Scanning Strategy
   
2. User Interaction with New Technology for Decision Making
   
3. Impact of DCAS Technology on Communication, Public Response, and Social and Economic Vulnerability
   
4. Establishing the Social and Economic Value of DCAS Information

Our strategy is to study and document user behavior, needs, and perceptions through in-depth interviews, surveys, quick response research, user participation in simulated weather scenarios, and user participation in CASA's test beds. We follow a spiral design process iterating between system feature development and demonstration/interaction with users.

Our research requires confronting CASA's dramatically new data paradigm for users, including user directed scanning of radars, dynamic data requests, high resolution weather data, and other changes to existing technology. Working closely with sophisticated data users, we employ primarily qualitative methodologies with a focus in individual perceptions, needs, decisions, and communications. To see the results of our research, please go to the End User Integration publications webpage.

Once the IP1 test bed is in full operation and users have the opportunity to use CASA data our focus will turn to validating and fine tuning our findings. Our research will expand to a wider range of users to understand group interactions with technology, decision making, societal response, and value through more quantitative studies.

Research Projects
Decision Sciences
The EUI Thrust is leading a new multi-disciplinary research effort to create an integrated decision model of CASA's end-to-end system (Integrated Systems Model) through a supplemental grant award. This Integrated Systems Model (ISM) will quantitatively link "upstream" technical capabilities, such as targeted radar observations, to their incremental impacts on later "downstream" responses such as warning decisions, risk communication, public response, and the resulting socio-economic impacts. The ISM will enable incorporation of socio-economic considerations into the end user policy and resource allocation algorithms, and also permit measurement of how different DCAS capabilities reduce negative socio-economic impacts.

In January 2007, a multidisciplinary group of CASA researchers representing all thrusts, ESEAB members, National Weather Service forecasters, Okalahoma Emergency Managers, and a television meteorologist began to define a detailed research plan. The ISM will model decisions related to severe thunderstorms, including hail and straight line winds, and tornadic events within the Oklahoma Test Bed between the time a watch is issued until an event or false alarm occurs. The research and validation of the model will occur through weather simulations at the Oklahoma Test Bed, and by leveraging existing EUI research on public response and operational forecaster decision making.

The ISM will consist of 4 mathematically linked sub-decision models: DCAS resource optimization, NWS warning, EM decisions, and public response related to impacts. For example, the output of the DCAS resource optimization sub-model would be the input to the NWS warning sub-model. The inputs and outputs to the sub-models are weather information: the attributes of weather data, such as feature strength, geographic specificity, or update time; and the quality/uncertainty of the data assessed by performance measures such as probability of detection (POD), false alarm rate (FAR), and lead time; and the manner in which the information is communicated. The supplement grant primarily funds the addition of decision scientist to the team, and expertise in the economic impacts and the measures that will link the models.

Simulations for End Users
This fall, we hosted a Weather Scenario Lab with Oklahoma Emergency Managers (EMs) who will participate in the Oklahoma test bed. These EMs used a CASA-developed tailorable scenario support tool. This support tool displayed, in simulated real time, historical radar data for a severe storm and the associated National Weather Service warning products. Two scenarios were shown: one with NEXRAD data, and another with NEXRAD data and CASA-like data. For each scenario, participants examined the information and discussed their weather assessments and associated actions they would take. The study provided additional insight into the decision making behaviors of EMs and how they might change with CASA's DCAS data, uncovered training and interface design issues, and illustrated how DCAS might be used in conjunction with other sources of information such as spotters.

Vulnerability Analysis
The EUI is focusing its research efforts on the determinants of vulnerability and how it impacts the severe weather decision-making process with regard to disaster preparedness and response. One project focuses on the development of a vulnerability index for the west coast of Puerto Rico integrating demographic and socio-economic data and climatic data to map public vulnerability to disaster in Puerto Rico. This information will be used to help determine radar placement in the Puerto Rico test bed and radar scanning strategies. Another project involves creating a integrated model of risk perception and protective action, focusing on public response to tornados warnings. Research includes in-depth interviews of the public following tornados in Louisiana, Missouri, Tennessee, and Oklahoma.

End User Policy Development
In the current system configuration, DCAS networks dynamically adjust their radar beams at 30 second intervals to sense the evolving weather optimally, feed data to customized weather detection and forecast algorithms and disseminate information to users based on their changing needs for data. For example, an NWS forecast office may use 360 degree sweeps of radar data close to the ground to monitor the evolution of a storm to determine whether to issue a tornado warning, while at the same time an emergency manager may need DCAS pinpointing capabilities, to locate precisely the most intense part of a storm for public notification or spotter deployment. These varying user information needs require different radar scanning strategies that, in some cases, exceed the resources of the DCAS networks. Which user information needs should the system serve first?

To address these potential resource conflicts, CASA has created an end user policy algorithm that maintains i) user rules, specifying in what manner and how often different kinds of weather phenomena should be scanned by radars and ii) user weights to establish the relative priority of different user groups in case of resource conflict. The end user policy algorithm interacts with the optimization and resource allocation algorithms that resolve resource conflicts and determine where the radars scan next. We have created the first version of end user policy based on qualitative input from expert users through in-depth interviews, review of best practices, and demonstrations of system output using simulations. We will conduct user experiments in the test beds and continue to evolve the end user policy.

Emergency Manager In-Depth Interviews
Interviews were conducted with representatives from Oklahoma's emergency management community and NWS meteorologists. These interviews have provided qualitative information for systems design on topics such as attitudes toward tornado tracking and detection, risk and frequency of weather-related threats, QPE, Emergency Manager priorities for end user policy, lead time perception, and have pointed to areas where additional research may be necessary.

(back to top)