The Water Institute is pleased to host a seminar by Dr. Lloyd A. Treinish titled:
Coupled environmental modelling for business decision making
Lloyd A. Treinish
Senior Technical Staff Member
Chief Scientist – IBM Deep Thunder Project
Smarter Planet – Industry Solutions
IBM Thomas J. Watson Research Center
Tuesday November 1, 2011
Centre for Environmental and Information Technology, Room 3142
1:30 pm
All are welcome.
Thank you.
Mary Anne Hardy
The Water Institute
University of Waterloo
519.888.4567 ext. 32658
mahardy@uwaterloo.ca
water.uwaterloo.ca
---- ABSTRACT ----
Water Institute Seminar
Tuesday, November 1, 2011, 1:30 pm
Centre for Environmental and
Information Technology (CEIT), Room 3142
University of Waterloo
Coupled Environmental Modelling for Business Decision Making
Dr. Lloyd Treinish
IBM Thomas J. Watson Research Center
Yorktown Heights, NY 10598
lloydt@us.ibm.com
The operation of any city, county or state is dependent to a significant degree upon weather conditions, especially with regard to relative extremes in wind, precipitation or temperature. For example, with precipitation events, local topography and weather conditions influence water runoff and infiltration, which directly affect flooding as well as drinking water quality and availability. The operation of the distribution system of an electric utility, particularly with an overhead infrastructure, can be highly sensitive to local weather conditions, including disruption by strong winds. The impact of such events creates issues of public safety for both citizens and first responders. Therefore, the availability of highly localized weather model predictions focused on municipal public safety operations has the potential to mitigate the impact of severe weather on citizens and local infrastructure. Typically, information at such a scale is simply not available. Hence, what optimization that is applied to these processes to enable proactive efforts utilizes either historical weather data as a predictor of trends or the results of synoptic-scale weather models. Neither source of information is appropriately matched to the temporal or spatial scale of many such operations. While near-real-time assessment of observations of current weather conditions may have the appropriate geographic locality, but by its very nature is only directly suitable for reactive response.
Current state-of-the-art numerical weather prediction (NWP) codes operating a the meso-gamma scale have been shown to have reasonable skill in being above to predict specific events or combination of weather conditions with sufficient spatial and temporal precision to address this scale mismatch. For example, we have been using the WRF-ARW (version 3.1.1) community NWP model to produce operational 84-hour forecasts for the New York City metropolitan area since early 2009, which are updated every twelve hours. It operates in a nested configuration, with the highest resolution at two km, utilizing 42 vertical levels. The configuration also includes parameterization and selection of physics options appropriate for the range of geography within the domain from highly urbanized to rural. This includes WSM-6 microphysics (includes explicit ice, snow and graupel), Yonsei University non-local-K scheme with explicit entrainment layer and parabolic K profile in the unstable mixed layer for the planetary boundary layer, Noah land-surface modeling with soil temperature and moisture in four layers, fractional snow cover and frozen soil physics, Grell-Devenyi ensemble cumulus parameterization, and the 3-category urban canopy model with surface effects for roofs, walls, and streets.
However, such a model-based weather forecast is only a prerequisite to the aforementioned optimization of weather-sensitive business operations. Based upon analysis of historical damage events and anecdotal evidence in the region, it is clear, for example, that storm-driven disruptions of the overhead electric distribution network (e.g., poles and wires) are caused by physical interaction of the atmosphere directly with that infrastructure or indirectly via nearby trees. However, reliable modelling with sufficient throughput for operational utilization of such interaction is neither tractable from a computational perspective nor verifiable from observations. Given that and the relative uncertainty in the processes and data, we approached the outage prediction from a stochastic perspective. Historical observations from a local, relatively dense mesonet operated by Earth Networks were analyzed along with data from a local electric utility concerning the characteristics of outage-related infrastructure damage from weather events, as well as the information about the distribution network and local environmental conditions. Data from this mesonet are also utilized to validate and improve the results of the WRF configuration. A Quasi-Poisson model was developed to statistically upscale the results of each WRF execution to irregularly shaped substation areas within the utility’s service territory. This has enabled the generation of forecasts of the number of jobs to be dispatched to repair storm-induced outages in each area. This approach also incorporates uncertainties in both the weather and outage data. Therefore, the model provides probabilities of outage restoration jobs per substation, which is associated with visualizations that explicitly depict the uncertainty. Based upon the analysis of the historical data, one of the key linkages between damage and severe storms is the magnitude of wind gusts. However, as is typical for NWP codes, WRF-ARW does not produce a direct representation of gusts. Therefore, a post-processing model was developed using a Bayesian hierarchical approach. It links the physical model outputs with real observations to create a calibrated, statistical representation of gusts. This model post-processes the sustained wind forecast on the topologically regular output generated by WRF based on the gust observations from the mesonet over time. It is used as one of the inputs to the storm impact for each substation area.
A prerequisite for a similar notion of a damage model for impacts due to flooding events requires an intermediate step. For example, the city of Rio de Janeiro, in Brazil, often faces the consequences of intense rainfall, which include landslides and flooding. In early April 2010, the city endured one of the worst torrential rainstorms in decades. At least 110 people were killed and tens of thousands lost their homes. To assist in planning for such events in the future, the city's leaders have enabled sophisticated capabilities for the coordinated management of disasters, emergencies, or planned events of importance. As part of that effort, the integration of advances in hydro-meteorological research is a key prerequisite. Given the geography of the city, such capabilities have significant challenges. In addition to its near-tropical setting along the coast of the Atlantic Ocean and the western portion of Guanabara Bay, there are regions where the terrain has a high aspect ratio, related to the Sierra do Mar mountains. Although sea breezes moderate the temperatures along the coast, especially during the summer, cold fronts from the Antarctic can lead to rapid changes in local weather. Of particular concern is the rainy summer season from December to March, during which O(100mm/day) precipitation events occur. Given the complex terrain and surface characteristics, significant flooding becomes likely during this period. Therefore, we have adapted WRF-ARW version 3.2.1 for use in the Rio de Janeiro area. An operational configuration was developed by retrospective analysis of recent significant precipitation events and compared against data from a network of 32 rain gauges operated by the city government during that period. Those results coupled with throughput considerations for availability of data for initial and boundary conditions as well as computational resources led to a configuration of four two-way nests focused on the Rio de Janeiro metropolitan area at 1km horizontal resolution. To address the influence of the complex terrain, 65 vertical levels were established with typically the lowest 15 or so being within the planetary boundary layer. The model orography was developed from altimetry data at 90m resolution available from the NASA Shuttle Radar Topography Mission. Data at 0.5 degree resolution from the NOAA Global Forecasting System are used for initial conditions and lateral boundaries. The configuration also has parameterization and selection of physics options appropriate for the range of geography in the region. It included the use of a sophisticated, double-moment, 6-class, explicit cloud microphysics scheme. This configuration was placed into operations in May 2011, producing 48-hour forecasts every 12 hours. The results of each model-based forecast are provided to the Government of Rio de Janeiro via a web portal in their integrated command center, which has been dubbed in Portuguese as Previsão Meteorológica de Alta Resolução (High-Resolution Weather Forecast) or PMAR. It includes HDTV-resolution animations of various two- and three-dimensional visualizations of key weather variables, specialized meteograms at locations of key landmarks, weather stations, etc. within the city and detailed tables of weather data at those locations. All of the web-based content contains information every 10 minutes of forecast time for each 48-hour model run. The visualizations are customized to the model configuration and the requirements of the end users, and incorporate data from the city’s geographic information system.
To further enable a prediction and warning system for weather impacts due to flooding, we have developed a hydrological model that operates at an urban scale. It employs a GIS-based unstructured two-dimensional mesh capable of capturing local terrain effects and simulating surface flow and water accumulation using a locally conservative approach. It is extended and simplified by employing one-dimensional mesh elements for street-level resolution that yields a reduction in computational cost without any loss of generality. The model is coupled to the aforementioned meteorological model or uses calibrated precipitation estimates derived from observing systems.
In both cases, in New York and Rio de Janerio, the coupled model approach has enabled prediction of storm impacts on local infrastructure as well as quantification of the uncertainty associated with such forecasts. We will discuss the approach and the background research effort, some specifics of how we brought these solutions into an operational phase, and lessons that were learned through the development and deployment. The work is on-going and the model results are being evaluated. We will present how the forecast information is being used and discuss the overall effectiveness of our approach for these and related applications as well as recommendations for future work.
--------------------------
Lloyd A. Treinish
Senior Technical Staff Member
Chief Scientist - Deep Thunder Project
Smarter Planet - Industry Solutions
IBM Thomas J. Watson Research Center
1101 Kitchawan Road, Yorktown Heights, NY 10598
No comments:
Post a Comment