Mission | People | Organization | Research | Education | Project | Data | Tools | Links
Water Quality and Wetland Assessment
Wetlands and Hydrology
Rainfall products for flood risk assessment in Virginia: Preliminary analysis of satellite and ground based rain productsProject Leader: Dr. Long. Chiu , GMU
The occurrence of floods is a major natural hazard that costs property damage and lives lost, including in Virginia. Floods occur due to an interaction between meteorological events and land surface properties. Rainfall statistics such as rainfall maximums, event total accumulations, and area rainfall are relevant to flood risk assessment. Because of the variability of space/time rainfall, these statistics are dependent on space/time sampling. In order to cover the space/time scale of flood risk analysis, spatially sparsely continuous gauge data have to be combined with intermittent satellite coverage. We use the following data sets for our analysis: 1) Fifty years of hourly precipitation gauge data (HPD) produced jointly by the National Climatic Data Center (NCDC) and the Forecast Systems Laboratory (FSL) from 38 gauges in Virginia; and 2) Fifty years of monthly 0.5 degree latitude/longitude land precipitation produced from gauge measurements by the University of Delaware. The satellite data consist of TRMM (Tropical Rainfall Measuring Mission) monthly 10 x 10 product (TSDIS referenced 3B43) based on satellite and NOAA operational gauge (CAMS) data and TRMM and other satellite daily 10 x 10 degree rainfall products (3B42).
We propose a comparison of space/time statistics, such as rain amount and frequency, conditional rain rate, derived from rain gauge data and satellite rain products derived from TRMM. Empirical Orthogonal Function (EOF) analysis will be performed on both satellite and gauge data sets to delineate the major spatial patterns of variability. Rain rate statistics for each hydrologic basin will be generated based on the EOF patterns. We also examine different variability timescales obtained from a time series power spectra for selected gauges.
Water Quality and Wetland Assessment
Project Leader: Drs. George Taylor (L) and David Wong, GMUBackground
A proposal was written in response to an EPA RFP for “Wetland Program Development Grants”. The competition was open to federal, state, local and non-profit groups with a goal of providing funds to improve/enhance wetland management programs. University researchers were eligible to apply as partners with federal, state or local agency or non-profit groups. Of the potential partners contacted, Fairfax County was most interested and quickly responded to the RFP request. The task associated with remote sensing supplement the EPA funded effort, enabling VAccess to mutually benefit from both EPA and NASA activities.The goal of this initiative is to focus on complementing the EPA and Fairfax County in one area in which the current effort is inadequate. This initiative is to evaluate the methodologies to evaluate changes over time in the health of wetlands at the county level. The traditional approach, Cross Correlation Analysis (CCA), is commonly used in regions of large-scale wetlands (e.g., coastal), but its reliability for urbanizing wetlands is suspect.
The objective is to investigate alternative “change detection” methodologies to assess the extent to which wetlands in Fairfax County have changed over time (last 2-2.5 decades) using a combination of remote sensing and ground-based data sets.
Technical Approach
Two major data inventories will be used and each is an independent assessment of the spatial extent of wetlands in Fairfax County, Virginia. The first is the NWI, (National Wetlands Inventory), an inventory conducted in the early 1980’s. The second is remotely sensed data from 2000 and 2001. Given the 20+ years between the two inventories, the objective is to compare the two databases and evaluate the degree of change spatially in the distribution of wetlands.
The effort will focus on 12 intensively inventoried, urbanizing wetlands in Fairfax County in the Piedmont region. These wetlands have been well mapped (2001 and 2002) using precise ecological (e.g., hydric soils) and jurisdictional features. This data set will be regarded as the basis for QA/QC. The 1980’s NWI and more recent remote sensing databases will be evaluated for change using statistical algorithms. The “proof of principle” will be based on change detection analysis for the 12 intensively studied wetland sites. Iterative methods will be employed to optimize the sensitivity of the statistical algorithms to detect change that is jurisdictionally and ecologically sound.
Milestones
Report on the change detection methodology as compared with other alternatives
Report using the change detection methodology at the county level
Flood Risk Analysis
Project Leader: Dr. Long Chiu, GMU
FLOODS
Background
The occurrence of flood is a major natural hazard that causes property damage and lost lives, representing a complex interaction among meteorology, hydrology and land surface interactions. While NOAA’s River Forecast Centers (RFC) provide large area flood warning, certain basins are especially prone to flood risk, simply because of their morphological characteristics. For example, the Tug Fork valley, McDowell County, West Virginia, experienced two flash floods within a period of 10 months. The recent May 3, 2002 flood resulted in 6 deaths. The prediction of flash flood in these small watersheds is especially challenging because of the size and network characteristics.
Technical Approach
To assess flash flood risk at roadway crossings for Fairfax county, we have taken three approaches. These approaches include visual inspection, rational method, and one based on morphometric parameters. The visual inspection method is based on in situ observation of flood debris remains, and is deemed incomplete. The rational method is based on historical rainfall measurements and hence the precipitation statistics need to be updated. We have developed algorithms to calculate morphometric parameters for water basins (reported earlier). These parameters are used for flood risk assessments at roadway crossings of watershed boundaries since the intersections represent risks at routes of evacuation. These algorithms have been applied to the Pond Branch Basin Watershed in northern Virginia. Morphometric parameters include drainage networks (such as total drainage segment number, total network length, network frequency, and density) and drainage basin characteristics (basin area, slope, perimeter, roughness, elongation, relief ratio, ruggedness number shape index aspect, circularity, etc). After the morphometric parameters are calculated for each basin, the flood risks for watersheds in the Pond Branch Basin watersheds are ranked. To fully assess the potential flood risks, statistics of precipitation, such as event maximum, duration, and extremes and social-economic data must also be integrated into a common database to support development design and emergency response decision-making.The objective of this task is to provide flood risk assessment of Virginia basins. To improve on our flood risk assessment, we will examine the following data sets and update our database for Virginia water basins.
1) The Shuttle Radar Topography Mission (SRTM) data set recently released by NASA. The data has 30m resolution and are degraded to 90m resolution for the higher level products. This data set will be used for morphometric analysis for Virginia basins.
2) USGS has also developed a hydrological database (HYDRO1K) at 1 km resolution that include DEM, derived flow direction, flow accumulation, slope and aspect. This data set will be compared with the SRTM data for consistency.
3) Hourly rain gauge data from NCDC. Hourly rain gauge data over 38 rain gauges have been acquired from NCDC. A preliminary assessment of the data quality (number of missing observations, duration) has been made. We plan to extend the coverage to include neighboring states and to compute rain fraction, conditional rain rates, event duration, intensity, and separation, and storm accumulation, maximum hourly and daily rain rate from these stations. These statistics will be ingested in Arc/View or Arc/GIS software to complement morphometric analysis
4) Hydrometeorological models for Flood forecasting. We will continue to evaluate coupled land surface-atmospheric models for flood forecasting and risk assessment. Potential candidates include the MM5, Mesoscale Compressible Community (MC2) coupled Canadian Land Surface Scheme (CLASS) model.
Milestones:
2nd Quarter Flood risk analysis assessments
WATER RESOURCESBackground:
Accurate precipitation estimates are important for water resource management. Rainfall accumulation and soil moisture information are useful for flood potential assessment.Technical Approach:
The study we propose for improving water resource management follows.
1) The water divisions of Virginia record monthly rain rate. This data is available for the past 130 years. We will examine seasonal, and interannaul variations of the rainfall and their relation to large-scale phenomena such as El Nino/Southern Oscillation (ENSO), the North Pacific (NPO), North Atlantic Oscillation (NAO), and Arctic Oscillation. Empirical Orthogonal Function (EOF) and time series (spectral, wavelet and Empirical Mode Decomposition) analyses will be performed on these data sets. Regression will be performed to examine cause and effect. We performed a preliminary analysis of the data. The results show little correlation between rainfall in Washington DC and the ENSO index for the 130 year time period, but the correlation is greatly enhanced in the past decade. Regression results will allow seasonal forecasts which may be of use to water managers.2) NEXRAD and TRMM data
We have contact with UNIDATA and plan to acquire realtime or near realtime NEXRAD radar data for regional flood risk monitoring. The NEXRAD radar data are available at roughly 10-minute intervals at 2 km resolution. A number of NEXRAD products are generated by UNIDATA. We will identify data types and coverage for our purpose. A problem of ground-based radar is beam blocking by mountains and between radar inter-calibration. A set of TRMM gridded space-borne Precipitation Radar (PR) data in GIS has been provided to us by the Goddard Space Flight Center Distributed Active Archive Center as part of the Remote Sensing Information Partner (RSIP) collaborative effort. These PR data will be used to compare gauge and NEXRAD data. We will also evaluate their use in regions where NEXRAD data is obstructed.3) Soil Moisture data.
We examined microwave signatures of soil moisture in lowly vegetated regions of the SW and found that there is significant signal from the TRMM Microwave Imager (TMI) data. Dr. Jackson of USDA is leading an effort to develop soil moisture data for the entire US. We contacted him and have established working relations with Dr. Jackson to examine the utility of the TRMM derived soil moisture data for the mid-Atlantic region. Soil moisture product will be available from the Advanced Microwave Scanning Radiometer (AMSR) on board the recently launched NASA AQUA satellite. We will evaluate the AMSR soil moisture when the data becomes available.Milestones:
3RD Quarter Flood potential analysis comparisons and assessments4th Quarter Final reports
VAccess-MAGIC © 2001-2002. All rights reserved
