H23E-1012
Snowmelt and Rainfall-Runoff Modeling Using NEXRAD Precipitation Input
Hydrologic simulation incorporating snowmelt dynamics is a challenging task many hydrologists face today. Since snow is composed of ice crystals and liquid water, the water equivalent of the snowpack is considered the primary contributor to streamflow generation in hydrologic dynamics, but the volume of meltwater varies depending on snow density and depth. Recently, the National Weather Service has also undertaken this challenge in order to enhance operational hydrologic prediction in Western basins in the United States. This research demonstrates how to integrate a NEXRAD data set, which is created using multi-sensor precipitation products, into the Hydrologic Simulation Program-Fortran (HSPF) and generate hydrologic runoff associated with the spatial precipitation retrieved from the NEXRAD data into a hydrologic modeling framework. The model performance of simulated streamflows was evaluated to assess the sensitivity of initial hydrologic parameters, including rain snow determination, Snow Water Equivalent (SWE), and other physical watershed configuration parameters. For snow simulation, the Temperature Index method was utilized to generate streamflow for the American and Carson basins, which are target basins of the Distributed Model Intercomparison Project (DMIP2). To measure simulation accuracy, evaluation statistics such as correlation coefficient, Nash-Sutcliffe efficiency, and exceedance probability between simulated and observed flow over calibration periods are also used. Additionally, through the framework of DMIP2, the model performance of simulated streamflow with calibration and without calibration was also evaluated to assess the sensitivity of initial parameters to calibrated parameters. Overall, calibrated simulation outperformed uncalibrated simulation for the western basins, and the results show that HSPF would be a suitable alternative model to reproduce the daily streamflows for many snow-driven watersheds.
H23E-1013
Radar Analyses of Space-Time Variability of Rainfall for Extreme Flood-Producing Storms in an Urban Environment
The Charlotte, North Carolina metropolitan area has experienced extensive urban and suburban growth during the past 40 years, resulting in increasing flood hazards in the region. Record flooding in the urban core of Charlotte occurred on 23 July 1997 from a storm that produced rainfall accumulations of more than 250 mm during an 18 hour period, more than doubling 24 hour accumulation maximum based on a 100 year rain gage record in Charlotte, and causing $60 million in property damage and three fatalities. Analyses of the 23 July 1997 storm and flood are based on rainfall and discharge observations from dense networks of rain gages and stream gages maintained by the U.S. Geological Survey and rainfall estimates from two WSR- 88D weather radars, both located approximately 100 km from the urban core of Charlotte. This wealth of observations provides an opportunity to address hydrometeorological questions concerning: (1) the accuracy of radar rainfall estimates for extreme, flood-producing rainfall; (2) the space-time variability of extreme, flood-producing rainfall in urban environments; (3) the effects of urbanization on extreme flood response in urban environments; and (4) the contrasts between extreme floods and more common floods. The authors show that bias-corrected radar rainfall estimates for the 23 July 1997 storm are quite accurate and provide the capability for resolving the fundamental rainfall forcing associated with regional variation in extreme flood response. This study shows how the spatial and temporal variability of rainfall combines with the land surface properties of the basins to determine the response of the different catchments to extreme rainfall in an urban setting.
H23E-1014
Evaluation of Hydro-NEXRAD Products in Southern California
The management of surface and groundwater water resources systems in semi-arid regions is a significant challenge. Accurate estimation of sparse and highly variable precipitation estimates is essential for hydrologic modeling of water supply, climate-related impacts, land-cover change studies and flood forecasting. This study compares rainfall estimates from ground-based rain gauges and radar-derived rainfall products in an attempt to assess the accuracy of radar rainfall estimates in Southern California. Twelve gauges in the Calleguas Creek watershed were analyzed for various storms in 2006-2007. The effectiveness of three different Hydro-NEXRAD algorithms (Quick Look, Pseudo NWS PPS, and Hi-Fi) was compared using the data from Sulphur Mountain NEXRAD located north of Los Angeles in Ventura County, California. Discrepancies between radar and gauge data were estimated in term of correlation coefficient, bias, Nash- Sutcliffe Efficiency (NSE) and covariance. As rainfall intensity increases, greater underestimation is observed in radar data. Relationships between radar accuracy and gauge elevation as well as distance from the radar station were examined. In general, NEXRAD radar estimates using the Quick Look algorithm had the best correlation with gauge values. The differences between the precipitation estimates of the various algorithms can be explained by the algorithm properties and methods used to derive estimates. A significant uncertainty in radar estimates is related to the ability of Sulphur Mountain NEXRAD to detect precipitation events below 1.83 km (6000 ft), which was also reported in other studies. Due to the lack of well-distributed ground-based systems in most regions in Southern California, bias-corrected radar estimates can provide improved estimates of precipitation fields enabling more accurate hydrologic modeling.
H23E-1015
Simulation of Urban Flooding in South-Central Texas
The San Antonio metropolitan area, is the 7th largest city in the nation, is located in one of the most flash- flood prone regions in North America and has experienced a number of flooding tragedies over the past decades. South Central Texas is particularly vulnerable to floods due to: (1) proximity to a moist air source (the Gulf of Mexico); (2) the Balcones Escarpment, which concentrates rainfall runoff; (3) a tendency for synoptic scale features to become cut-off and stall over the area; and (4) decaying tropical cyclones stalling over the area. This presentation will discuss physically based distributed parameters hydrologic modeling studies in the region. Predicted runoff from a number of catchments ranging in size from less that 50 km2 to over 3000 km2 is compared to observations. Radar precipitation was used as input. Flood events simulated include the 2002, 2004, and 2007 events. Particular focus is on flooding events in San Antonio and Austin areas. Impact of the characteristics of inputs and watershed physiography on model predictions will be discussed.
H23E-1016
NEXRAD Radar-based Hydraulic Flood Prediction System for a Major Evacuation Route in Houston
With the advent of GIS, radar-based rainfall estimation using NEXRAD, and delivery systems on the internet, flood warning systems can provide communities and major traffic roads with important advanced warning of impending flood conditions. A major northbound evacuation route from Gulf Coast crossing Brays Bayou has historically been inundated due to its low elevation of the road deck. In order to relieve flood damages and save lives, a real-time NEXRAD radar-based flood alert system was developed for this cross section with an in-depth investigation focusing on the downstream section of Brays Bayou. This new system was developed utilizing existing hydrologic/hydraulic models, NEXRAD radar data, and the framework of the existing flood alert system (FAS2).The development followed three major steps: 1) creation of a real-time hydrologic model (RTHEC-1) from the current HEC-HMS model for Brays Bayou to accurately predict peak flows at the bridge; 2) addition of a predictive FloodPlain Map Library (FPML) to the current FAS2 to delineate floodplains and water surface elevations under various spatial and temporal conditions associated with rainfall intensities; and, 3) development of a centralized monitoring system incorporating local rain gauges and stream gauges. The FPML module is able to analyze NEXRAD radar rainfall intensities and patterns to quantify water surface elevations and delineate floodplains in real time, and will enable emergency personnel to begin flood preparation with as much lead time as possible.
H23E-1017
A GIS Tool for evaluating and improving NEXRAD and its application in distributed hydrologic modeling
In this study, a user friendly GIS tool was developed for evaluating and improving NEXRAD using raingauge data. This GIS tool can automatically read in raingauge and NEXRAD data, evaluate the accuracy of NEXRAD for each time unit, implement several geostatistical methods to improve the accuracy of NEXRAD through raingauge data, and output spatial precipitation map for distributed hydrologic model. The geostatistical methods incorporated in this tool include Simple Kriging with varying local means, Kriging with External Drift, Regression Kriging, Co-Kriging, and a new geostatistical method that was newly developed by Li et al. (2008). This tool was applied in two test watersheds at hourly and daily temporal scale. The preliminary cross-validation results show that incorporating raingauge data to calibrate NEXRAD can pronouncedly change the spatial pattern of NEXRAD and improve its accuracy. Using different geostatistical methods, the GIS tool was applied to produce long term precipitation input for a distributed hydrologic model ˇ§C Soil and Water Assessment Tool (SWAT). Animated video was generated to vividly illustrate the effect of using different precipitation input data on distributed hydrologic modeling. Currently, this GIS tool is developed as an extension of SWAT, which is used as water quantity and quality modeling tool by USDA and EPA. The flexible module based design of this tool also makes it easy to be adapted for other hydrologic models for hydrological modeling and water resources management.
H23E-1018
Hydrometeorological Analysis of Flooding Events in San Antonio, TX
South Central Texas is particularly vulnerable to floods due to: proximity to a moist air source (the Gulf of Mexico); the Balcones Escarpment, which concentrates rainfall runoff; a tendency for synoptic scale features to become cut-off and stall over the area; and decaying tropical cyclones stalling over the area. The San Antonio Metropolitan Area is the 7th largest city in the nation, one of the most flash-flood prone regions in North America, and has experienced a number of flooding events in the last decade (1998, 2002, 2004, and 2007). Research is being conducted to characterize the meteorological conditions that lead to these events and apply the rainfall and watershed characteristics data to recreate the runoff events using a two- dimensional, physically-based, distributed-parameter hydrologic model. The physically based, distributed–parameter Gridded Surface Subsurface Hydrologic Analysis (GSSHA) hydrological model was used for simulating the watershed response to these storm events. Finally observed discharges were compared to GSSHA model discharges for these storm events. Analysis of the some of these events will be presented.
H23E-1019
Flood and inundation evaluation using a GIS based hydrological model
In this paper, the SOBEK model was used which combines the function of a one-dimensional channel flow and a two-dimensional overland flow to analyze and present the flood potential of the Hsin-Huwei Creek basin in Taiwan. In the article, the methodology illustrated involved integrating the DTM data and hydrologic data, calibrating both hydrologic and hydraulic models, and producing inundation maps directly usable for planning of flood-prone areas. The GIS was used to perform the tedious and time-consuming tasks of spatial analysis for the 5 m × 5 m resolution DTM data. Using real rainfall data with observed channel stages, the parameters of the model were calibrated. The model was finally used for inundation simulation, and the results were used for inspecting the flood control facilities and drainage systems to reduce flood threats. Keywords Inundation evaluation, GIS, Hydrological model
H23E-1020
Early Evaluation of NEXRAD Super-Resolution Precipitation Estimates for Hydrologic Applications
Recently, NEXRAD's WSR-88D weather radars commenced providing enhanced base data that have higher resolution, called super-resolution, than the legacy-resolution of a 1 km by 1 degree polar grid. The super- resolution data are produced for "split cuts" that depend on Volume Coverage Pattern and the grid spacing of reflectivity data are reduced to 0.5 degree in azimuth and to 250 m in range. Although super-resolution data may capture small scale features of precipitation with more reliability, NEXRAD Precipitation Processing System still operates based on recombined (legacy-resolution) data. The algorithms to quantitatively estimate precipitation from super-resolution data involve three components: 1) preprocessing, 2) rain rate, and 3) rainfall accumulation. Preprocessing algorithm performs super-resolution volume scan data quality control and builds "hybrid scan" that is, super-resolution reflectivity values for each azimuth and range bin are assigned from the several lowest elevation angles. It also corrects range effects using azimuth dependent vertical reflectivity profile. Rain rate algorithm converts the corrected super-resolution reflectivity data to rainfall intensity by specifying the power-law type empirical relationship between reflectivity and rainfall intensity. Rainfall accumulation algorithm integrates consecutive rain rate scans for specific time duration ranging from 5 minutes to daily. Since super-resolution data have been implemented only in recent months, the authors present algorithm testing and comparison results of the super-resolution precipitation estimates for just few rain events that led to extreme flooding in the Iowa and Cedar River basins. They also evaluate the super-resolution precipitation estimates with recombined legacy products and rain gauge data.
H23E-1021
Real Time Detection of Anomalies in Streaming Radar and Rain Gauge Data
Radar-rainfall data are being used in an increasing number of real-time applications because of their wide spatial and temporal coverage. Because of uncertainties in radar measurements and the relationship between radar measurements and rainfall on the ground, radar-rainfall data are often combined with rain gauge data to improve their accuracy. While rain gauges can provide accurate estimates of rainfall, their data are sometimes subject to a number of errors caused by the environment in which the gauges are deployed. This study develops a method for automatically detecting anomalies (i.e. data that deviate markedly from historical patterns) in both radar and raingauge data through integration and modeling of data from these two different sources.. These anomalous data can be caused by sensor or data transmission errors or by infrequent system behaviors that may be of interest to the scientific or public safety communities. This study develops an automated anomaly detection method that employs a Dynamic Bayesian Network to assimilate data from multiple rain gauges and weather radar (NEXRAD) into an uncertain model of the current rainfall. Filtering (e.g. Kalman filtering) can then be used to infer the likelihood that a particular gauge measurement is anomalous. Measurements with a high likelihood of being anomalous are classified as such. The method developed in this study performs fast, incremental evaluation of data as they become available; scales to large quantities of data; and requires no a priori information regarding process variables or types of anomalies that may be encountered. The performance of the anomaly detector developed in this study is demonstrated using a precipitation sensor network composed of a NEXRAD weather radar and several near- real-time telemetered rain gauges deployed by the USGS in Chicago. The results indicate that the method performs well at identifying anomalous data caused by a real sensor failure.
H23E-1022
Geo-Spatial Grid-based Transformations of Multi-Sensor Precipitation Estimates
Geo-spatial interpolation methods are often necessary in instances where the precipitation estimates available from multi-sensor or NEXRAD for a specific spatial grid needs to be transformed to another grid with a different spatial grid or orientation. The study current focuses on development, evaluation and implementation of spatial interpolation or weighting methods for transforming multi-sensor estimate based precipitation data available in the form of 4km x 4 km HRAP (hydrologic rainfall analysis project) grid to a 2 km x 2 km NEXRAD grid. Six weighting methods are developed to achieve this geometric transformation of precipitation estimates in space and time. The methods use distances and aerial extents of intersected segments of the grids as weights in the interpolation schemes and were applied to transform MPE to NEXRAD grid in South Florida Water Management District (SFWMD) region. A total of 192 rain gages in this region were used as ground truth to assess the quality of precipitation estimates obtained via these interpolation methods and are also used for bias correction procedures. To help in the assessment, several error measures are calculated and appropriate weighting functions are developed to select the best method suitable for the transformation.
H23E-1023
Comparative Evaluation of NEXRAD and TRMM Satellite based Precipitation Estimates and Rain Gage Measurements
The availability of accurate precipitation data is essential for hydrologic modeling and water resources management. Recent advances NEXRAD and satellite based measurements have provided alternative methods for precipitation estimation. The current study focuses on comparative evaluation and assessment of TRMM (Tropical Rainfall Measuring Mission) satellite based precipitation estimates. Multi-sensor estimate based precipitation data available in the form of 4km x 4 km HRAP (hydrologic rainfall analysis project) and 2 km x 2 km NEXRAD grid based precipitation estimates, and rain gage measurements available in a specific region of South Florida region were used for comparison with the TRMM based precipitation measurements. The utility of satellite-based precipitation data is also evaluated for infilling missing precipitation data at rain gages. A total of 58 rain gages in this region were used as ground truth to assess the quality of precipitation estimates. Preliminary results suggest that TRMM satellite based precipitation estimates compare well with NEXRAD and rain gage measurements.
H23E-1024
Analaysis of San Antonio River Floods Caused by Tropical Storm Erin
Tropical Storm Erin started as a depression on August 14 2007. It deepened rapidly to evolve into a tropical storm the morning of the 15th. It moved into Texas on the 16th with maximum sustained winds of 56 km/hr. The storm produced 2-10 inches over south central Texas on August 16-17, 2008. The heaviest rainfall fell within a 6-hour period with totals in excess of 7.5 inches, as observed by the WSR-88D radar in New Braunfels, near San Antonio, TX. Average precipitation over the summer provided sufficient moisture to cause Erin's precipitation to produce significant rapid runoff over portions of the San Antonio River. Radar rainfall data and a two-dimensional, physically-based, distributed-parameter hydrologic model were used to perform hydrometeorological analysis of this event. Hydrologic simulations on several sub-basins will be discussed.
H23E-1025
Radar Detected Rainfall Intensity As An Input For Shallow Landslides Slope Stability Model
The term "shallow landslides" is widely used in literature to describe a slope movement of limited size that mainly develops in soils up to a maximum of a few meters. Shallow landslides are usually triggered by heavy rainfall because, as the water starts to infiltrate in the soil, the pore-water pressure increases so that the shear strength of the soil is reduced leading to slope failure. For this work we have developed a distributed hydrological-geotechnical model for the forecasting of the temporal and spatial distribution of shallow landslide to be used as a warning system for civil protection purpose. The main goal of this work is the use of radar detected rainfall intensity as the input for the hydrological simulation of the infiltration. Using the rainfall pattern detected by the radar is in fact possible to dynamically control the redistribution of groundwater pressure associated with transient infiltration of rain so as to infer the slope stability of the studied area. The model deals with both saturated and unsaturated conditions. Two pilot sites have been chosen to develop and test this model: the Armea basin (Liguria, Italy) and the Ischia Island (Campania, Italy). In recent years several severe rainstorms have occurred in both these areas. In at least two cases these have triggered numerous shallow landslides that have caused victims and damaged roads, buildings and agricultural activities. In its current stage the basic basin-scale model applied for predicting the probable location of shallow landslides involves several stand-alone components. A module for estimating the groundwater pressure head distribution according to radar detected rainfall intensity, a soil depth prediction scheme and a limit-equilibrium infinite slope stability algorithm which produces a factor of safety (FS). The additional ancillary data required have been collected during the field work. The single components are seamlessly integrated into a system that automatically publishes constantly updated FS values to a WebGIS in near-real- time so that local administrators responsible for public safety can access and download the data from the internet. This system has been running for a few months and is now being validated. Several types of problems hinder a correct validation of the system. One major obstacle was overcome when major storms triggered several tens of soil slips in December 2006 for the Armea basin and in April 2006 for Ischia. This events provided both the necessary rainfall data for the soil saturation component, which until then for previous occurred landslides was lacking, and a new landslide inventory for comparison with the FS produced by the slope stability model for the same event. The inventory was derived from a newly acquired VHR satellite image. Another important aspect of the research being performed regards the assessment of the relative importance of the different parameters involved in the limit-equilibrium infinite slope stability model. This statistical sensitivity analysis has the aim of determining which errors in the input variables slope gradient, soil depth, soil saturation, cohesion and angle of internal friction produce the largest errors in the output FS values. Preliminary results indicate the importance of topographic attributes and of soil depth.
H23E-1026
Application of Volumetric Weather Radar Data and the Distributed Rainfall Runoff Model REW in the Ourthe Catchment
In the southern Ardennes region of Belgium near the border with Luxembourg, the Royal Meteorological Institute of Belgium (RMI) installed a C-band Doppler weather radar at an elevation of 600 m in the year 2001. This volumetric weather radar scans over multiple elevations at a temporal resolution of 5 minutes. The current study explores the possibility of using the volumetric information of the precipitation field to correct for the effects of the Vertical Profile of Reflectivity (VPR) over the period October 1, 2002 until March 31, 2003. During this winter half year storm events are mainly stratiform, giving rise to bright band effects which can decrease the performance of the radar. Previous studies have shown multiple drawbacks in applying a single estimated VPR profile to correct such reflectivity data. Therefore, the focus here is on the temporal variability of the VPR as measured by the radar and its variability over different spatial scales. This information is applied to generate a number of possible rainfall fields. These realizations are employed to try to quantify some of the discrepancies in precipitation intensities as estimated by the weather radar and those measured by a raingauge network. The final step then is to assess their potential within a distributed rainfall runoff model. The 1597 km2 Ourthe catchment lies within 60 km of the radar. Over this medium sized watershed ten raingauges measuring at an hourly interval are more or less equally distributed. Near the outlet discharge data are collected at the same time step. The distributed hydrological Representative Elementary Watershed (REW) model is applied to model the hydrological behavior of the Ourthe over the six month period. The benefits of the high spatial and temporal resolution of weather radar data compared to a conventional raingauge network plus the possibility of generating multiple realizations of the precipitation field are expected to yield more information about the hydrological behavior of the Ourthe catchment.
H23E-1027
On hydrologic prediction using NEXRAD Radar for recharge estimation
Advances in water balance and recharge estimation have been aided through radar detection of complex precipitation patterns and distributed hydrologic modeling. Predicting the hydrologic components of the water balance depends on accurate determination of precipitation input, runoff, infiltration, recharge, and evapotranspiration. The importance of gauge-corrected radar input is demonstrated through characterizing the climatology at basin and subbasin scales. The difference in runoff volume obtained from corrected and uncorrected radar input is significant even at length scales on the order of 100's of km. Bias correction of the NEXRAD precipitation estimates from show that there is a decline in prediction accuracy. Random errors that remain after bias correction are expected to cancel out over time; however, uncorrected radar introduces error in the hydrologic water balance. Actual ET is an important factor affecting recharge in the upper reaches of the Blue River and the Arbuckle-Simpson aquifer. Characterizing local precipitation rates becomes a critical determinant in understanding recharge rates to the aquifer, runoff, and springflow.
H23E-1028
Multi-Sources Precipitation Estimates
Development of a multi-source merging approach to improve Satellite-based Precipitation Estimates (SPE), particularly for areas that radar network cannot cover is the objective of this study. Merging multi sources rainfall such as SPE, ground-based radar (RR), and rain gauge (RG) measurements is a technique for reducing the impacts of various uncertainties on generated rainfall. Enhancement of remotely sensed rainfall estimates is very important because satellite is the only possible source of capturing information from the areas where radar and gauge network cannot cover. Although, satellite is a source of collecting information with no spatial limitation, precipitation estimates from satellite imagery have greater uncertainties particularly on estimating precipitation intensity. Hence, application of remote sensing data for precipitation estimation, particularly over the remote and mountainous regions, where there is usually heavier precipitation and cannot be completely covered by ground-based observation sources, is a challenging research area. In this study, a weighting approach, Successive Correction Method (SCM), will be used to merge different gridded data. Kriging technique is used to interpolate and convert point rain-gauge observations into spatially aggregated areal measurements. Point rain-gage measurements have been changed to spatial measurements of 4km resolution using Kriging with point to point scheme. Satellite-based NESDIS rainfall algorithm, Hydro-Estimator (HE) is selected to be merged with rain gauge- and radar-based data. Preliminary results of merging SPE and radar using SCM algorithm demonstrated that the SCM technique is capable of extending radar and information from pixels with available radar and/or rain gauge rainfall to their neighboring pixels with no ground-based information. The first step in this approach is bias correction of SPE with respect to radar and/or rain gauge rainfall data and then applying the merging algorithm to combine multi-sensor rainfall information.
H23E-1029
Polarimetric X-band radar network simulator to analyze attenuation correction algorithms
Radars operating at X-band frequencies often need to have correction algorithms applied to the
measurements to correct for attenuation of radar signals by the intervening rainfall. Existing attenuation
correction algorithms applicable to weather radars are based on polarimetric measurements such as specific
differential phase and differential reflectivity. One of the disadvantages of existing algorithms is the instability
in case when total signal attenuation is observed.
In this work, authors present an approach to attenuation correction techniques utilizing a network of radars
with overlapping view of a domain of interest, e.g. a small watershed or a city. To assess performance of the
proposed method, the authors developed a full radar network simulator that works with output rainfall fields
from a numerical mesoscale model. The simulator is based on the T-matrix approach and provides
polarimetric variables for individual radars constituting the network. Developed simulator is used to assess
attenuation effects and evaluate different attenuation correction techniques and strategies. Obtained results
will be used to support optimal configuration of an operational radar network. Early simulation results are
presented and discussed.