Radar Rainfall Estimates Research Articles

Effects of Rain Gauge Temporal Resolution on the Specification of a Z–R Relationship

Abstract The weather radar is an efficient alternative for measuring spatially varying rainfall covering a large area at a high temporal resolution. This paper studies the impact of rainfall gauge temporal resolution on optimal relationships between radar reflectivity (Z) and rainfall rate (R). Four datasets of radar reflectivity and corresponding rain gauge rainfall data from Sydney and Brisbane, Australia, and one dataset from Bangkok, Thailand, were used in the analysis. Climatological Z–R relationships were calibrated using rainfall aggregated over 1–24 h to investigate the evidence of temporal scaling in the Z–R calibrated parameters. This analysis points to an increase in the multiplicative term (the A parameter) of the Z–R relationship as temporal resolutions become finer. This pattern is repeated in all the datasets analyzed. Thereafter, a simple scaling hypothesis was proposed to develop transformations that could scale the A parameter in the Z–R relation across a range of temporal resolutions. This scaling relationship was found to be suitable, with the scaling exponent attaining values close to 0.055 across all the datasets analyzed. The proposed relationship has a significant role in radar rainfall estimation studies, especially in regions where subdaily gauge rainfall measurements are not readily available to ascertain optimal Z–R parameters.

Bollène-2002 Experiment: Radar Quantitative Precipitation Estimation in the Cévennes–Vivarais Region, France

Abstract The Bollène-2002 Experiment was aimed at developing the use of a radar volume-scanning strategy for conducting radar rainfall estimations in the mountainous regions of France. A developmental radar processing system, called Traitements Régionalisés et Adaptatifs de Données Radar pour l’Hydrologie (Regionalized and Adaptive Radar Data Processing for Hydrological Applications), has been built and several algorithms were specifically produced as part of this project. These algorithms include 1) a clutter identification technique based on the pulse-to-pulse variability of reflectivity Z for noncoherent radar, 2) a coupled procedure for determining a rain partition between convective and widespread rainfall R and the associated normalized vertical profiles of reflectivity, and 3) a method for calculating reflectivity at ground level from reflectivities measured aloft. Several radar processing strategies, including nonadaptive, time-adaptive, and space–time-adaptive variants, have been implemented to assess the performance of these new algorithms. Reference rainfall data were derived from a careful analysis of rain gauge datasets furnished by the Cévennes–Vivarais Mediterranean Hydrometeorological Observatory. The assessment criteria for five intense and long-lasting Mediterranean rain events have proven that good quantitative precipitation estimates can be obtained from radar data alone within 100-km range by using well-sited, well-maintained radar systems and sophisticated, physically based data-processing systems. The basic requirements entail performing accurate electronic calibration and stability verification, determining the radar detection domain, achieving efficient clutter elimination, and capturing the vertical structure(s) of reflectivity for the target event. Radar performance was shown to depend on type of rainfall, with better results obtained with deep convective rain systems (Nash coefficients of roughly 0.90 for point radar–rain gauge comparisons at the event time step), as opposed to shallow convective and frontal rain systems (Nash coefficients in the 0.6–0.8 range). In comparison with time-adaptive strategies, the space–time-adaptive strategy yields a very significant reduction in the radar–rain gauge bias while the level of scatter remains basically unchanged. Because the Z–R relationships have not been optimized in this study, results are attributed to an improved processing of spatial variations in the vertical profile of reflectivity. The two main recommendations for future work consist of adapting the rain separation method for radar network operations and documenting Z–R relationships conditional on rainfall type.

New paradigm for statistical validation of satellite precipitation estimates: Application to a large sample of the TMPA 0.25° 3‐hourly estimates over Oklahoma

It is well acknowledged that space‐based and ground‐based radar estimates of rainfall are both affected by several sources of uncertainties; nevertheless, the latter are generally used as ground reference for the validation of the former. In this way, it is difficult to assess what portion of the observed discrepancies is solely due to the targeted satellite errors and what is the impact of the radar‐rainfall errors. The authors propose a novel approach for the validation of satellite‐based rainfall estimates, in which an empirically based radar‐rainfall error model is used to account for deterministic and random uncertainties in radar‐rainfall estimates. Using an easy‐to‐interpret metric, the approach can be used to answer probabilistic questions such as “Given radar and satellite rainfall estimates, what is the probability that the true rainfall is larger (smaller) than the satellite estimate”? The results are based on a large sample (6 years) of the Tropical Rainfall Measuring Mission (TRMM) Multisatellite Precipitation Analysis (TMPA) 0.25‐degree 3‐hourly satellite‐rainfall estimates and hourly radar‐rainfall product (aggregated at the satellite space‐time scale) from the Oklahoma City WSR‐88D radar (KTLX). This study shows how the TMPA satellite product tends to overestimate the high rainfall values and underestimate the low values and how it tends to perform better during the hot season (June to September). Moreover, the authors demonstrate how the presence of errors in the ground‐based radar could significantly affect the results of the satellite product evaluation.

A Study of the Error Covariance Matrix of Radar Rainfall Estimates in Stratiform Rain. Part II: Scale Dependence

Abstract The contribution of various physical sources of uncertainty affecting radar rainfall estimates at the ground has been recently quantified at a resolution typically used in schemes assimilating rainfall at the ground onto mesoscale models. Here, the contribution of the two most important sources of uncertainty at nonattenuating wavelengths (the range-dependent error and the uncertainty due to the Z–R transformation) and their interaction are studied as a function of the resolution of radar observations. The analysis is carried out using a large dataset of collocated reflectivity profiles from the McGill S-band radar and disdrometric measurements obtained in stratiform rainfall at resolutions of 1 × 1, 5 × 5, and 15 × 15 km2. Results show that the errors affecting radar quantitative precipitation estimation (QPE) have a strong dependence with range, and that their structure is scale dependent. At the analyzed resolutions, QPE errors are significantly correlated in time and over several grid points.

Hydrometeorological modelling for flash flood areas: the case of the 2002 Gard event in France

AbstractIn the context of flash flood forecasting, this paper proposes a few advances in our understanding of the hydrometeorological processes and their associated modelling requirements that may be useful to introduce within an operational forecasting chain. The study is focused on the September 2002 storm that produced more than 600 mm of rainfall in <24 h and triggered a series of flash floods in the South of France. This catastrophic event took 23 human lives in 16 distinct subcatchments. This paper proposes a combined detailed analysis of the meteorological event and hydrological simulations of the response of four small‐ungauged catchments. The meteorological analyses are based on observations and results of simulation of rain fields obtained with the MesoNH model. These analyses explained the steadiness of the storms that led to a locally intense precipitation: the role of the orography and favourable synoptic conditions. The hydrological model is set up without any calibration and the soil parameter specification is based on an existing soil database. Radar rainfall estimations are used. Simulated specific peak discharges are found to be in agreement with estimations from a postevent in situ investigation. Based on the model results, a cartography of the dominant process is proposed for the four selected catchments.

Comparing the Rainfall Patterns Produced by Hurricanes Frances (2004) and Jeanne (2004) over Florida

During 2004, Hurricanes Frances and Jeanne produced heavy rainfall across Florida. Although making landfall at the same location and moving along similar tracks, these storms produced different rainfall patterns. We employ a GIS to analyze rain gauge and radar-estimated rainfall data spatially, and we determine which physical mechanisms caused the asymmetrical rain swaths produced by each storm. We find that due to a moist environment and weakening vertical wind shear, the rain fields of Frances covered a large area. Strong convection in several outer rainbands and slow forward velocity yielded high rainfall totals. Although Jeanne contained more convective rainfall than Frances, vertical wind shear, a faster forward velocity, and existence within a relatively dry environment caused rainfall totals to be lower. The techniques employed in this study could be applied to a larger number of landfalling tropical cyclones to better understand where high rain totals occur under different combinations of environmental conditions.

The Effect of Storm Life Cycle on Satellite Rainfall Estimation Error

Abstract The study uses storm tracking information to evaluate error statistics of satellite rain estimation at different maturity stages of storm life cycles. Two satellite rain retrieval products are used for this purpose: (i) NASA’s Multisatellite Precipitation Analysis–Real Time product available at 25-km/hourly resolution (3B41-RT) and (ii) the University of California (Irvine) Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks (PERSIANN) product available at 4-km–hourly resolution. Both algorithms use geostationary satellite infrared (IR) observations calibrated to an array of passive microwave (PM) earth-orbiting satellite sensor rain retrievals. The techniques differ in terms of algorithmic structure and in the way they use the PM rainfall to calibrate the IR rain algorithms. The satellite retrievals are evaluated against rain gauge–calibrated radar rainfall estimates over the continental United States. Error statistics of hourly rain volumes are determined separately for thunderstorm and shower-type convective systems and for different storm life durations and stages of maturity. The authors show distinct differences between the two satellite retrieval error characteristics. The most notable difference is the strong storm life cycle dependence of 3B41-RT relative to the nearly independent PERSIANN behavior. Another is in the algorithm performance between thunderstorms and showers; 3B41-RT exhibits significant bias increase at longer storm life durations. PERSIANN exhibits consistently improved correlations relative to the 3B41-RT for all storm life durations and maturity stages. The findings of this study support the hypothesis that incorporating cloud type information into the retrieval (done by the PERSIANN algorithm) can help improve the satellite retrieval accuracy.

A modeling approach to assess the hydrological response of small mediterranean catchments to the variability of soil characteristics in a context of extreme events

Abstract. This paper presents a modeling study aiming at quantifying the possible impact of soil characteristics on the hydrological response of small ungauged catchments in a context of extreme events. The study focuses on the September 2002 event in the Gard region (South-Eastern France), which led to catastrophic flash-floods. The proposed modeling approach is able to take into account rainfall variability and soil profiles variability. Its spatial discretization is determined using Digital Elevation Model (DEM) and a soil map. The model computes infiltration, ponding and vertical soil water distribution, as well as river discharge. In order to be applicable to ungauged catchments, the model is set up without any calibration and the soil parameter specification is based on an existing soil database. The model verification is based on a regional evaluation using 17 estimated discharges obtained from an extensive post-flood investigation. Thus, this approach provides a spatial view of the hydrological response across a large range of scales. To perform the simulations, radar rainfall estimations are used at a 1 km2 and 5 min resolution. To specify the soil hydraulic properties, two types of pedotransfer function (PTF) are compared. It is shown that the PTF including information about soil structure reflects better the spatial variability that can be encountered in the field. The study is focused on four small ungauged catchments of less than 10 km2, which experienced casualties. Simulated specific peak discharges are found to be in agreement with estimations from a post-event in situ investigation. Examining the dynamics of simulated infiltration and saturation degrees, two different behaviors are shown which correspond to different runoff production mechanisms that could be encountered within catchments of less than 10 km2. They produce simulated runoff coefficients that evolve in time and highlight the variability of the infiltration capacity of the various soil types. Therefore, we propose a cartography distinguishing between areas prone to saturation excess and areas prone only to infiltration excess mechanisms. The questions raised by this modeling study will be useful to improve field observations, aiming at better understanding runoff generation for these extreme events and examine the possibility for early warning, even in very small ungauged catchments.

Distributed hydrological modelling using weather radar in gauged and ungauged basins

Distributed hydrological modelling using space–time estimates of rainfall from weather radar provides a natural approach to area-wide flood forecasting and warning at any location, whether gauged or ungauged. However, radar estimates of rainfall may lack consistent, quantitative accuracy. Also, the formulation of hydrological models in distributed form may be problematic due to process complexity and scaling issues. Here, the aim is to first explore ways of improving radar rainfall accuracy through combination with raingauge network data via integrated multiquadric methods. When the resulting gridded rainfall estimates are employed as input to hydrological models, the simulated river flows show marked improvements when compared to using radar data alone. Secondly, simple forms of physical–conceptual distributed hydrological model are considered, capable of exploiting spatial datasets on topography and, where necessary, land-cover, soil and geology properties. The simplest Grid-to-Grid model uses only digital terrain data to delineate flow pathways and to control runoff production, the latter by invoking a probability-distributed relation linking terrain slope to soil absorption capacity. Model performance is assessed over nested river basins in northwest England, employing a lumped model as a reference. When the distributed model is used with the gridded radar-based rainfall estimators, it shows particular benefits for forecasting at ungauged locations.

Validation of TRMM precipitation radar monthly rainfall estimates over Brazil

In an attempt to validate the Tropical Rainfall Measuring Mission (TRMM) precipitation radar (PR) over Brazil, TRMM PR estimates are compared with rain gauge station data from Agência Nacional de Energia Elétrica (ANEEL). The analysis is conducted on a seasonal basis and considers five geographic regions with different precipitation regimes. The results showed that TRMM PR seasonal rainfall is well correlated with ANEEL rainfall (correlation coefficients are significant at the 99% confidence level) over most of Brazil. The random and systematic errors of TRMM PR are sensitive to seasonal and regional differences. During December to February and March to May, TRMM PR rainfall is reliable over Brazil. In June to August (September to November) TRMM PR estimates are only reliable in the Amazonian and southern (Amazonian and southeastern) regions. In the other regions the relative RMS errors are larger than 50%, indicating that the random errors are high.

On the propagation of uncertainty in weather radar estimates of rainfall through hydrological models

AbstractThe generation of flow forecasts using rainfall inputs to hydrological models has been developed over many years. Unfortunately, errors in input data to models may vary considerably depending upon the different sources of data such as raingauges, radar and high resolution Numerical Weather Prediction (NWP) models. This has hampered the operational use of radar for quantitative flow forecasting. The manner with which radar rainfall input and model parametric uncertainty influence the character of the flow simulation uncertainty in hydrological models has been investigated by several authors. In this paper an approach to this problem based upon a stochastic hydrological model is considered. The errors in the input data, although they may be constrained, do propagate through the model to the flow predictions. Previous work on this error propagation through a fully distributed model is described, and a similar analysis for a stochastic hydrological model implemented in a mixed rural and urban area in north‐west England carried out. Results are compared with those previously published for an American catchment. A possible approach to selecting flow forecast ensemble members is proposed. Copyright © 2009 Royal Meteorological Society

A Comparison of Several 5-Minute Radar—Rainfall Estimation Models

For the Z-R relationship in radar—based rainfall estimation, the distribution of corresponding R values for a given Z value (or the corresponding Z value for a given R value) may be highly skewed. However, the traditional power—law model is physically deduced and fitted under the normal—distribution presumption of radar wave echoes associated with a rain rate value, and it may not be very appropriate. Considering this problem, the authors devised several generalized linear models with different forms and distribution presumptions to represent the Z-R relationship. Radar—reflectivity scans observed by a CINRAD/SC Doppler radar and 5-minute rainfall accumulation recorded by 10 ground gauges were used to fit these models. All data used in this study were collected during some large rainfalls of the period from 2005 to 2007. The radar and all gauges were installed in the catchment of the Yishu River, a branch of the Huaihe River in China. Three models based on normal distribution and a dBZ presumption of gamma distribution were fitted using maximum—likelihood techniques, which were resolved by genetic algorithms. Comparisons of estimated maximized likelihoods based on assumptions of gamma and normal distribution showed that all generalized linear models (GLMs) of presumed gamma distribution were better fitted than GLMs based on normal distribution. In a comparison of maximum—likelihood, the differences between these three models were small. Three error statistics were used to assess the agreement between radar estimated rainfall and gauge rainfall: relative bias (B), root mean square error (RMSE), and correlation coefficient (r). The results showed that no one model was excellent in all criteria. On the whole, the GLM—based models gave smaller relative bias than the traditional power—law model. It is suggested that validations conducted in many previous works should have been made against a specific criterion but overlooked others.

Application of Weather Radar in Estimation of Bulk Atmospheric Deposition of Total Phosphorus Over Lake Simcoe

The decline of Lake Simcoe water quality has been attributed to high phosphorus inputs that result in excessive algae and macrophyte growth subsequently contributing to end-of-summer hypolimnetic dissolved oxygen depletion and loss of fish habitat. Out of the estimated 53 to 67 tonne/annum (1998 to 2004 water years) of phosphorus entering the lake, atmospheric deposition is believed to be responsible for 16 to 38 tonne/annum. Historical estimates for atmospheric deposition involved averaging rain gauge (rainfall depth) and rain quality (phosphorus concentration) station data. Through use of this procedure, any spatial variability in the data (quality and quantity) is lost as each gauge is given an equal weighting. This study proposes a methodology to use Next Generation Radar (NEXRAD) to spatially represent rainfall data and a method to correct radar-rainfall estimates to rainfall recorded by local rain gauges. From this analysis it was found that the radar generally represented localized rainfall well, with the majority of correlation coefficients (R2) being over 0.90. Radar related issues that resulted in poor R2 values included virga, overshooting beam, beam attenuation, range related issues and ground clutter. For large bulk atmospheric total phosphorus (TP) deposition events the dominant parameter in calculating TP loads was rainfall depth. Results from this analysis demonstrated a large (−88% to +44%) difference between historical and revised estimates of bulk atmospheric deposition of phosphorus over Lake Simcoe.

A Study of the Error Covariance Matrix of Radar Rainfall Estimates in Stratiform Rain

Abstract The contribution of various physical sources of uncertainty affecting radar rainfall estimates at the ground is quantified toward deriving and understanding the error covariance matrix of these estimates. The focus here is on stratiform precipitation at a resolution of 15 km, which is most relevant for data assimilation onto mesoscale numerical models. In the characterization of the error structure, the following contributions are considered: (i) the individual effect of the range-dependent error (associated with beam broadening and increasing height of radar measurements with range), (ii) the error associated with the transformation from reflectivity to rain rate due to the variability of drop size distributions, and (iii) the interaction of the first two, that is, the term resulting from the cross correlation between the effects of the range-dependent error and the uncertainty related to the variability of drop size distributions (DSDs). For this purpose a large database of S-band radar observations at short range (where reflectivity near the ground is measured and the beam is narrow) is used to characterize the range-dependent error within a simulation framework, and disdrometric measurements collocated with the radar data are used to assess the impact of the variability of DSDs. It is noted that these two sources of error are well correlated in the vicinity of the melting layer as result of the physical processes that determine the density of snow (e.g., riming), which affect both the DSD variability and the vertical profile of reflectivity.

The Effect of Clustering on the Uncertainty of Differential Reflectivity Measurements

Abstract One of the most important avenues of recent meteorological radar research is the application of polarization techniques to improve radar rainfall estimation. A keystone in many of these methods is the so-called differential reflectivity ZDR, the ratio of the reflectivity factor ZH at horizontal polarization backscattered from a horizontally polarized transmission to that corresponding to a vertically polarized transmission ZV. For such quantitative applications, it is important to understand the statistical accuracy of observations of ZDR. The underlying assumption of all past estimations of meteorological radar uncertainties is that the signals obey Rayleigh statistics. It is now evident, however, that as a radar scans, the meteorological conditions no longer always satisfy the requirements for Rayleigh statistics. In this work, ZDR is reconsidered, but this time within the new framework of non-Rayleigh signal statistics. Using Monte Carlo experiments, it is found that clustering of the scatterers multiplies the standard deviation of ZDR beyond what is always calculated assuming Rayleigh statistics. The magnitude of this enhancement depends on the magnitudes of the clustering index and of the cross correlation between ZH and ZV. Also, it does not depend upon the number of independent samples in an ensemble estimate. An example using real radar data in convective showers suggests that non-Rayleigh signal statistics should be taken into account in future implementations of polarization radar rainfall estimation techniques using ZDR. At the very least, it is time to begin to document the prevalence and magnitude of the clustering index in a wide variety of meteorological conditions.

An integrated model for predicting rainfall-induced landslides

This study proposes a novel method that combines a deterministic slope stability model and a statistical model for predicting rainfall-induced landslides. The method first uses the deterministic model to derive the rainfall rate critical to induce slope failure for each land unit. Then it calculates the difference between the critical rainfall threshold and estimated rainfall intensity. Using the difference and estimated rainfall duration as explanatory variables, the method derives a logit (integrated) model to compute landslide occurrence probabilities. To demonstrate the effectiveness of this method, the study used radar rainfall estimates and landslides associated with a typhoon (tropical cyclone) to develop the integrated model and the same types of data associated with another typhoon to validate the model. The model had a modified success rate of 84.0% for predicting landslides and stable areas, and model validation yielded a modified success rate of 87.4%. Both rates were better than those from the critical rainfall model. The main advantage of the integrated model lies in its use of rainfall variables that are not included in calculating the critical rainfall. Also, as a probabilistic model, the integrated model is better suited for decision-making in watershed management. This study has advanced the method for predicting rainfall-triggered landslides.

An operational approach for classifying storms in real-time radar rainfall estimation

Summary This paper presents an operational approach to integrate a storm classification method into a real-time radar rainfall estimation. A minor modification of a texturing classification algorithm proposed by Steiner et al. [Steiner, M., Houze Jr., R.A., Yuter, S.E., 1995. Climatological characterisation of three-dimensional storm structure from operational radar and rain gauge data. J. Appl. Meteorol. 34, 1978–2007] that can classify each pixel in the radar image as stratiform or convective is used to classify the instantaneous reflectivity field into convective and stratiform components. A method to derive the climatological Z–R relations for convective and stratiform rainfall is proposed. Vertical profiles of reflectivity (VPRs) are used to verify the accuracy of the storm classification. An alternative method for verification of a storm classification scheme based on differences between probability distribution functions of rain gauge rainfall of the two rainfall categories is also presented. A 6-month record of radar and rain gauge rainfall for Sydney, Australia for November 2000–April 2001 is used for training and rainfall events during February 2007–March 2007 are used to evaluate the efficiency and applicability of the proposed methods. The results show that the proposed operational approach for classifying storms in real-time radar rainfall estimates reduces RMSE between radar and rain gauge rainfall from 4.63 to 4.30 mm h −1 and from 5.31 to 5.01 mm h −1 for the training and verification periods compared to the case where the rainfall is assumed to have a single Z–R relationship.

Estimation of Rainfall Based on the Results of Polarimetric Echo Classification

Abstract The quality of polarimetric radar rainfall estimation is investigated for a broad range of distances from the polarimetric prototype of the Weather Surveillance Radar-1988 Doppler (WSR-88D). The results of polarimetric echo classification have been integrated into the study to investigate the performance of radar rainfall estimation contingent on hydrometeor type. A new method for rainfall estimation that capitalizes on the results of polarimetric echo classification (EC method) is suggested. According to the EC method, polarimetric rainfall relations are utilized if the radar resolution volume is filled with rain (or rain and hail), and multiple R(Z) relations are used for different types of frozen hydrometeors. The intercept parameters in the R(Z) relations for each class are determined empirically from comparisons with gauges. It is shown that the EC method exhibits better performance than the conventional WSR-88D algorithm with a reduction by a factor of 1.5–2 in the rms error of 1-h rainfall estimates up to distances of 150 km from the radar.

Modeling radar-rainfall estimation uncertainties using parametric and non-parametric approaches

There are large uncertainties associated with radar estimates of rainfall, including systematic errors as well as the random effects from several sources. This study focuses on the modeling of the systematic error component, which can be described mathematically in terms of a conditional expectation function. The authors present two different approaches: non-parametric (kernel-based) and parametric (copula-based). A large sample (more than six years) of rain gauge measurements from a dense network located in south-west England is used as an approximation of the true ground rainfall. These data are complemented with rainfall estimates by a C-band weather radar located at Wardon Hill, which is about 40 km from the catchment. The authors compare the results obtained using the parametric and non-parametric schemes for four temporal scales of hydrologic interest (5 and 15 min, hourly and three-hourly) by means of several different performance indices and discuss the strengths and weaknesses of each approach.

Application of radar data to modeling rainfall-induced landslides

Many shallow landslides are triggered by heavy rainfall. Previous studies of landslide modeling have been hampered by the scarcity of rain gauges to adequately depict the spatial variability of rainfall conditions triggering landslides. This study simulates the efficiency of the critical rainfall model for landslide prediction in a mountainous watershed with inputs of different rainfall estimates associated with a typhoon (tropical cyclone) event. Doppler radar data at a spatial resolution of 1 km and measurements from six gauging stations provide the sources for rainfall estimates. Inverse distance weighted, splines, and kriging are the interpolation methods for gauged rainfall estimates. A comparison of the simulation outputs shows that the model using radar rainfall estimates has a better performance than those using gauged rainfall estimates in predicting both landslides and stable areas. In light of the results, this paper also discusses the validity of the critical rainfall model for landslides in relation to its rainfall input and steady-state hydrological assumption.