In this paper, we present a comprehensive review of the data sources and estimation methods of 30 currently available global precipitation data sets, including gauge‐based, satellite‐related, and reanalysis data sets. We analyzed the discrepancies between the data sets from daily to annual timescales and found large differences in both the magnitude and the variability of precipitation estimates. The magnitude of annual precipitation estimates over global land deviated by as much as 300 mm/yr among the products. Reanalysis data sets had a larger degree of variability than the other types of data sets. The degree of variability in precipitation estimates also varied by region. Large differences in annual and seasonal estimates were found in tropical oceans, complex mountain areas, northern Africa, and some high‐latitude regions. Overall, the variability associated with extreme precipitation estimates was slightly greater at lower latitudes than at higher latitudes. The reliability of precipitation data sets is mainly limited by the number and spatial coverage of surface stations, the satellite algorithms, and the data assimilation models. The inconsistencies described limit the capability of the products for climate monitoring, attribution, and model validation.

An artificial neural network (ANN) is a flexible mathematical structure which is capable of identifying complex nonlinear relationships between input and output data sets. ANN models have been found useful and efficient, particularly in problems for which the characteristics of the processes are difficult to describe using physical equations. This study presents a new procedure (entitled linear least squares simplex, or LLSSIM) for identifying the structure and parameters of three-layer feed forward ANN models and demonstrates the potential of such models for simulating the nonlinear hydrologic behavior of watersheds. The nonlinear ANN model approach is shown to provide a better representation of the rainfall-runoff relationship of the medium-size Leaf River basin near Collins, Mississippi, than the linear ARMAX (autoregressive moving average with exogenous inputs) time series approach or the conceptual SAC-SMA (Sacramento soil moisture accounting) model. Because the ANN approach presented here does not provide models that have physically realistic components and parameters, it is by no means a substitute for conceptual watershed modeling. However, the ANN approach does provide a viable and effective alternative to the ARMAX time series approach for developing input-output simulation and forecasting models in situations that do not require modeling of the internal structure of the watershed. IntroductionThe rainfall-runoff (R-R) process is believed to be highly nonlinear, time-varying, spatially distributed, and not easily described by simple models [Singh, 1964; Kulandais•amy and Subramanian, 1967;Chiu and Huang, 1970; Pilgrim, 1976]. Two major approaches for modeling the R-R process have been explored in the literature: "conceptual" (physical) modeling and "system theoretic" modeling (sometimes referred to as "black box") [Amorocho and Hart, 1964;Sorooshian, 1983; Guoeta, 1984;Singh, 1988;Duan, 1991; O'Connell and Clarke, 1981; Young and Wallis, 1985]. Conceptual rainfall-runoff (CRR) models are designed to approximate within their structures (in some physically realistic manner) the general internal subprocesses and physical mechanisms which govern the hydrologic cycle. CRR models usually incorporate simplified forms of physical laws and are generally nonlinear, time-invariant, and deterministic, with parameters that are representative of watershed characteristics. Until recently, for practical reasons (data availability, calibration problems, etc.) most conceptual watershed models assumed lumped representations of the parameters. Among the more widely used and reported lumped parameter watershed models are the Sacramento soil moisture accounting (SAC-SMA) model of the U.S. National Weather Service [Burnash et al., 1973; Brazil and Hudlow, 1980], HEC-1 [U.S. Army Corps of Engineers, 1990], and the Stanford watershed model (SWM) [Crawford and Linsley, 1966]. While such models ignore the spatially distributed, time-varying, and stochastic properties of the R-R process, they attempt to incorporate realistic rep...

A new retrospective satellite-based precipitation dataset is constructed as a climate data record for hydrological and climate studies. Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks–Climate Data Record (PERSIANN-CDR) provides daily and 0.25° rainfall estimates for the latitude band 60°S–60°N for the period of 1 January 1983 to 31 December 2012 (delayed present). PERSIANN-CDR is aimed at addressing the need for a consistent, long-term, high-resolution, and global precipitation dataset for studying the changes and trends in daily precipitation, especially extreme precipitation events, due to climate change and natural variability. PERSIANN-CDR is generated from the PERSIANN algorithm using GridSat-B1 infrared data. It is adjusted using the Global Precipitation Climatology Project (GPCP) monthly product to maintain consistency of the two datasets at 2.5° monthly scale throughout the entire record. Three case studies for testing the efficacy of the dataset against available observations and satellite products are reported. The verification study over Hurricane Katrina (2005) shows that PERSIANN-CDR has good agreement with the stage IV radar data, noting that PERSIANN-CDR has more complete spatial coverage than the radar data. In addition, the comparison of PERSIANN-CDR against gauge observations during the 1986 Sydney flood in Australia reaffirms the capability of PERSIANN-CDR to provide reasonably accurate rainfall estimates. Moreover, the probability density function (PDF) of PERSIANN-CDR over the contiguous United States exhibits good agreement with the PDFs of the Climate Prediction Center (CPC) gridded gauge data and the Tropical Rainfall Measuring Mission (TRMM) Multi-Satellite Precipitation Analysis (TMPA) product. The results indicate high potential for using PERSIANN-CDR for long-term hydroclimate studies in regional and global scales.

PERSIANN, an automated system for Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks, has been developed for the estimation of rainfall from geosynchronous satellite longwave infared imagery (GOES-IR) at a resolution of 0.25° x 0.25° every half-hour. The accuracy of the rainfall product is improved by adaptively adjusting the network parameters using the instantaneous rain-rate estimates from the Tropical Rainfall Measurement Mission (TRMM) microwave imager (TMI product 2A12), and the random errors are further reduced by accumulation to a resolution of 1° x 1° daily. The authors' current GOES-IR-TRMM TMI based product, named PERSIANN-GT, was evaluated over the region 30°S-30°N, 90°E-30°W, which includes the tropical Pacific Ocean and parts of Asia, Australia, and the Americas. The resulting rain-rate estimates agree well with the National Climatic Data Center radar-gauge composite data over Florida and Texas (correlation coefficient p > 0.7). The product also compares well (p ~ 0.77-0.90) with the monthly World Meteorological Organization gauge measurements for 5° x 5° grid locations having high gauge densities. The PERSIANN-GT product was evaluated further by comparing it with current TRMM products (3A11, 3B31, 3B42, 3B43) over the entire study region. The estimates compare well with the TRMM 3B43 1° x 1° monthly product, but the PERSIANN-GT products indicate higher rainfall over the western Pacific Ocean when compared to the adjusted geosynchronous precipitation index-based TRMM 3B42 product.