H52A-01 INVITED
Snowflake Size Distribution Measurements in South Central Ontario, Canada
In support of NASA's Global Precipitation Measurement (GPM) mission ground validation program, NASA's two laser optical disdrometers (Parsivel) and Colodaro State University (CSU) two-dimensional video disdrometer (2dvd) were deployed to a well-equipped precipitation observation site in South Central Ontario, Canada. The instruments were collocated and have been operating since late November 2006. So far, there has been numerous lake effect and synoptic winter storms over the site. In one event, parsivel disdrometers recorded 50 cm of snowfall. In addition, there have been at least 10 storms where the snow accumulation exceeded 4 cm. The leading objective of this study was to compare the parsivel and 2dvd size and fall velocity measurements for selected cases and relate the findings to the physical processes within and below the cloud. Unlike 2dvd, parsivel measures the maximum dimension of the snowflake in a single plane, while the fall velocity is calculated from the duration of the flake within the laser beam. The 2dvd samples the same flake in two planes from which fall velocity is obtained. The 2dvd also measures the maximum width and height in both planes. At the time of this abstract, two parsivels and 2dvd were operated nearly continuously for almost three months and preliminary data analysis is encouraging. The field site, which is known as Centre for Atmospheric Research Experiments (CARE), is an atmospheric research facility operated by the Air Quality Research Branch of the Meteorological Service of Canada and is located 80 km north of Toronto, Ontario, Canada in a rural agricultural and forested region. During the past three winters, a field campaign was conduced in support of Canadian CloudSat/CALIPSO validation project (c3vp). However, 2006-07 winter was the first since the satellites were in orbit. The coordinated efforts of aircraft missions over the CARE facility during the Intensive Operation Periods will enhance our understanding of cold cloud microphysics.
H52A-02 INVITED
Satellite Rainfall Estimates And Prediction Of Extreme Precipitation Events Over The Mexico City Basin.
Precipitation estimation schemes using remote sensors in satellites have been developed over the last three decades. The most interesting fact of the Tropical Rainfall Measuring Mission (TRMM) platform is the conjunction of passive and active sensors in order to obtain precipitation estimates over the ocean and remote continental regions. Comparisons of rainfall estimates for the Mexico City metropolitan area were made with data interpolated from TRMM and a dense rain gauge network. Results show good agreement between the data sets for both spatial distribution of rain and rain intensity estimations. The estimates have been used to test the ability of a mesoscale weather prediction model to predict extreme events based on a simple 1-2-3 rule. Results show that the model underestimates amounts of intense precipitation events, except when convective activity is organized by synoptic scale circulations such as easterly waves.
H52A-03
Properties of Small-Scale Rainfall Spatial Correlation in Central Oklahoma
Spatial variability of rainfall in the range of small scales between the size of a single raingauge (of the order of 0.1 m) and a few kilometers is an important problem from practical and as well as scientific points of view. For example, it is a source of fundamental difficulties in meaningful evaluation of radar rainfall estimates. Mathematical models of rainfall morphology have the capability to describe a broad span of spatiotemporal scales, however, their development and validation require adequate empirical evidence. Most of the experimental studies on spatial rainfall structure are based on the weather radar observations that are inherently averaged over areas ranging from about 4 km2 to 25 km2, and cannot provide information on finer scales. Therefore, spatial rainfall variability at the scales below a few kilometers is still a poorly explored research area. In this study, we present an extensive analysis of the small-scale variability in rainfall fields that is focused on their spatial correlation structure at the distances below 3-4 km. The correlation functions are well established, normalized and commonly used measures of spatial dependences that are required by many applications. In contrast to the multi-fractal measures, they can be estimated based on a network of point rainfall measurements. The correlation estimates presented here are based on a large data sample from a unique local cluster of 25 stations, called the EVAC PicoNet, covering an area of about 9 km2 in Central Oklahoma. High density of the network and relatively large size of the sample allowed us to look at the small-scale rainfield correlations from three different perspectives. First, we examined the dependences of the spatial correlation function on the time- scale for the averaging intervals spanning the range from 1 minute to 1 day. Second, we present and analyze the differences between the correlation estimates in the individual rainfall events. Finally, we demonstrate and discuss ambiguities associated with the conditioning of the correlation estimates on rainfall intensities.
H52A-04
Multi-sensor QPE in the National Mosaic and QPE (NMQ) System
Accurate quantitative precipitation estimation (QPE) and forecast are critical for flood and flash flood warnings and
for water resource managements. Significant advancements in recent years in computational resources,
networking, and remote sensing technologies have provided great opportunities to develop more accurate QPE
than what were there before. Given the complex spatial and temporal characteristics of precipitation processes,
not one single observing system can provide complete and accurate measurements of surface precipitation for
the wide spectrum of hydrological applications. For instance, the rain gauges make direct measurements of the
surface precipitation, but the gauge stations are often sparsely distributed and the measurements are subject to
errors due to temporary blockage of the collecting orifice by frozen hydrometeors, wind blowing-off effects on the
tipping buckets, telemetry errors, etc. Weather radars make semi-direct measurements of the precipitation but
there are uncertainties associated with reflectivity (Z) - rain rate (R) conversion in addition to errors associated
with beam blockage and non-uniform vertical profile of reflectivity. Satellite data have no obstructions and provide
best coverage of precipitation systems among all observational networks, yet they only observe cloud tops and
provide indirect measurements of precipitation. The multi-sensor QPE in the NMQ system makes use of
advantages of each observing systems and produces integrated precipitation products. The radar data in the
optimal sampling area are used to dynamically calibrate the satellite infrared field and to produce a satellite QPE
product. The satellite QPE is combined with radar-based QPE to fill in regions with poor radar coverage and a
blended radar-satellite QPE is generated. The rain gauge data are then used to adjust the magnitude of the
radar-satellite blended QPE field and to remove bias. This paper presents an overview of the NMQ multi-sensor
QPE schemes currently executing nationally in real time and some initial performance results.
http:www.nmq.nssl.noaa.gov
H52A-05
Radar Observations of Rainfall Variability Using Non-Rayleigh Signal Fluctuations
The spatial and temporal variability of precipitation is widely recognized. In particular, rainfall rates (R) fluctuate the most in regions where the raindrops are clustered and where mean conditions are changing (statistical heterogeneity). For purposes requiring as quantitative radar measurements as possible, this variability may often challenge our ability to make precise measurements. Indeed, at times the very meaning of an average rainfall rate (whether using conventional or polarimetric radars) is likely to become quite ambiguous. It would, therefore, be useful to identify those locations where this variability may be particularly problematic. In this work a technique is proposed and applied to quantify this variability using deviations from Rayleigh statistics of intensity (I) measurements. Using results from a different paper in this conference, it is shown analytically that in conditions of clustering and statistical heterogeneity, the square of the intrinsic relative dispersion of Z is the clustering index for each component of the heterogeneous rain appropriately weighted by the square of the fractional contribution that each component makes to the reflectivity factor. A technique is described for separating the Rayleigh signal contributions to the observed relative dispersions from those arising from clustering and statistical heterogeneities. Applications to conventional meteorological radar measurements are illustrated. Often, but not always, the greatest ambiguities in estimates of the average rainfall rate occur just where R are the largest and presumably where accurate estimates are most important. This ambiguity is not statistical. Instead these results show precisely where no single average value applies uniformly to the entire domain. These examples also demonstrate that the appropriate observations are feasible using current conventional meteorological radars with adequate processing capabilities. However, changes in radar technology necessary to make such observations more routinely and more precisely are also suggested.
H52A-06
Hydrologic Modeling of the Lower Los Amigos Amazon Watershed in Southeastern Peru with MIKE-SHE using TRMM Precipitation Data
An integrated 3-hourly step time surface-subsurface hydrologic model for the Lower Los Amigos watershed was generated using the MIKE-SHE model. Weather data was obtained from a local observation station, and the TRMM satellite. We compare and contrast modeling using local observations vs. satellite rainfall data, and modeling limitations imposed by remotenes of study area.
H52A-07
Advances in Precipitation Forecast Verification: The Forecast Quality Index and a Case Study in Assessing WRF Model Performance
The problem of developing metrics for high-resolution forecast verification which are capable of capturing the essential multiscale space-time structure of observed and modeled fields and can also provide constructive feedback to model developers, has gained renewed interest in the atmospheric and hydrometeorology communities as they have realized that current indices are inadequate. Recently, we proposed a novel measure called Forecast Quality Index (FQI) for QPF verification, based on the classical Hausdorff distance, the Universal Image Quality Index (an image similarity measure used in image processing) and the concept of "surrogate" images. Comparing with the traditional measures on several simulated test images and real precipitation datasets (from the Storm and Mesoscale Ensemble Experiment - SAMEX'98) we demonstrated the potential of this newly proposed measure. In this presentation, we will elaborate on the performance of FQI in the Spatial Verification Methods Inter-comparison Project (organized by the Research Applications Laboratory, NCAR). As part of this inter-comparison project, several cases from the 2005 SPC/NSSL Spring Experiment were selected. NCEP Stage II hourly precipitation analyses and several high-resolution Weather Research and Forecasting (WRF) model forecasts were used as observation-forecast pairs. The results will be placed in the context of other competing metrics.
H52A-08
Hurricanes, Extreme Rainfall, and Subseasonal Predictability for the US, Mexico, and Central America
The influence of hurricanes on extreme daily rainfall events and the potential subseasonal predictability of those events due to the influence of the Madden-Julian Oscillation (MJO) are examined for the US, Mexico, and Central America. More than 15,000 daily precipitation stations from the Global Daily Climatology Network (GDCN) are used to analyze extreme rainfall events. Hurricane and tropical storm locations are taken from the Hurricane Best Track (HURDAT) files of the National Hurricane Center for the Atlantic and Eastern North Pacific, 1974-2000. The Wheeler and Hendon real-time index is used to characterize the MJO activity and the Maharaj and Wheeler approach is used to forecast the MJO. As expected, hurricane activity strongly influences the occurrence of extreme daily rainfall events over large areas of the region and the daily station data provides the maximum resolution of the signal, especially in regions with sparse data coverage. Of particular note are the large spatial scales of the influence of hurricanes in their extratropical stage over land, resulting in the hurricane influence extending considerably further inland than the well-known coastal impacts. Given the previously-established MJO influence on hurricane activity and the link between hurricanes and extreme rainfall events, the MJO influence on extreme rainfall events is examined both for hurricane-related events and for non-hurricane-related events. The MJO influence on daily rainfall extremes in Mexico identified in our previous work is seen to be closely related to the MJO modulation of hurricane activity. The influence of hurricanes and the MJO on flood disasters in Central America and Mexico is examined in terms of societal impact using the EM-DAT disaster database of the Centre for Research on the Epidemiology of Disasters (CRED). Since the MJO is predictable at subseasonal timescales, the MJO influence on extreme events (both hurricane and non-hurricane related) may also be predictable and this is examined at the station level based on the Maharaj and Wheeler MJO forecast.