Downscaled Precipitation Data Research Articles

Spatial downscaling analysis of GPM IMERG precipitation dataset based on multiscale geographically weighted regression model: a case study of the Inner Mongolia Reach of the Yellow River basin

The Inner Mongolia Reach of the Yellow River Basin is characterized by a relative scarcity of meteorological stations. While satellite precipitation products can complement observations from meteorological stations, their limited spatial resolution restricts their efficacy in regional studies. This study utilizes the GPM IMERG precipitation dataset, considering various factors that influence the spatial distribution of precipitation, such as the Normalized Difference Vegetation Index (NDVI), elevation, slope, aspect, and topographical relief, to construct a multiscale geographically weighted regression (MGWR) model. A spatial downscaling method for the GPM IMERG precipitation dataset is proposed, and its reliability is validated through an accuracy assessment. Moreover, the scale differences in the impact of different factors on the spatial pattern of precipitation in the Inner Mongolia Reach of the Yellow River Basin are scrutinized. The results indicate that: 1) The downscaled GPM IMERG precipitation data (1 km × 1 km) exhibit enhanced accuracy compared to the pre-downscaled data (approximately 11 km × 11 km). The correlation coefficient, Bias, and RMSE of the annual precipitation data after downscaling of GPM IMERG are 0.865, 6.05%, and 68.50 mm/year, respectively. For the monthly downscaled precipitation data, the correlation coefficient, Bias, and RMSE are 0.895, 6.09%, and 16.25 mm/month, respectively. The downscaled GPM IMERG precipitation dataset exhibit high accuracy on both annual and monthly temporal scales. 2) Different factors demonstrate localized effects on precipitation in both dry and wet years. Elevation is the dominant factor influencing the spatial heterogeneity of annual precipitation. The findings from this study can provide technical support for hydrological modeling, drought monitoring, and water resource management in data-scarce areas of the Inner Mongolia Reach of the Yellow River Basin.

Development of flood forecasting and warning system using hybrid approach of ensemble and hydrological model for Dharoi Dam

Abstract The most frequent natural disaster is flooding. Advanced forecasting systems are lacking in developing countries. The majority of urban areas are located close to flood plains for rivers. Accurate flood forecasting is necessary for reservoir planning and flood management. The Sabarmati River's atmospheric-hydrologic ensemble flood forecasting model has been developed using TIGGE data. Precipitation can be reliably predicted by TIGGE's global ensemble numerical weather prediction (NWP) systems. By using NWP data, flood forecasting systems may be extended from hours to days. Ensemble weather forecasts are produced using the European Centre for Medium-Range Weather Forecasts and National Centers for Environmental Prediction together with 5-day lead times from TIGGE. The flood occurrences from 2015, 2017, and 2020 were used for the calibration and validation of the ensemble flood forecasting model. Bias was corrected using Bayesian model averaging (BMA), heterogeneous extended linear regression, censored non-homogeneous linear regression (cNLR), and other statistical downscaling techniques. Forecasted and downscaled precipitation data were checked using the Brier score and rank likelihood score. For cNLR, Brier's score performed admirably. The specificity vs. sensitivity performance of the cNLR and BMA approaches is 91.87 and 91.82%, respectively, according to receiver operating characteristic and area under the curve diagrams. Models with the hybrid hydrologic coupling approach accurately predict floods. Users may predict peak time and peak discharge hazard likelihood with reliability using peak time and flood warning probability distributions.

SpateGAN: Spatio‐Temporal Downscaling of Rainfall Fields Using a cGAN Approach

AbstractClimate models face limitations in their ability to accurately represent highly variable atmospheric phenomena. To resolve fine‐scale physical processes, allowing for local impact assessments, downscaling techniques are essential. We propose spateGAN, a novel approach for spatio‐temporal downscaling of precipitation data using conditional generative adversarial networks. Our method is based on a video super‐resolution approach and trained on 10 years of country‐wide radar observations for Germany. It simultaneously increases the spatial and temporal resolution of coarsened precipitation observations from 32 to 2 km and from 1 hr to 10 min. Our experiments indicate that the ensembles of generated temporally consistent rainfall fields are in high agreement with the observational data. Spatial structures with plausible advection were accurately generated. Compared to trilinear interpolation and a classical convolutional neural network, the generative model reconstructs the resolution‐dependent extreme value distribution with high skill. It showed a high fractions skill score of 0.6 (spatio‐temporal scale: 32 km and 1 hr) for rainfall intensities over 15 mm h−1 and a low relative bias of 3.35%. A power spectrum analysis confirmed that the probabilistic downscaling ability of our model further increased its skill. We observed that neural network predictions may be interspersed by recurrent structures not related to rainfall climatology, which should be a known issue for future studies. We were able to mitigate them by using an appropriate model architecture and model selection process. Our findings suggest that spateGAN offers the potential to complement and further advance the development of climate model downscaling techniques, due to its performance and computational efficiency.

Review on spatial downscaling of satellite derived precipitation estimates

The present work aims at reviewing and identifying gaps in knowledge and future perspectives of satellite-derived precipitation downscaling algorithms. Here, various aspects related to statistical and dynamical downscaling approaches of the precipitation data sets from the Tropical Rainfall Measuring Mission (TRMM) and its successor Intergraded Multi-Satellite Retrievals for Global Precipitation Measurement (IMERG–GPM) mission are reviewed and the existing downscaling methods are categorized and analysed, to highlight the usefulness and applicability of the produced downscaled precipitation data sets. In addition, a critical comparison of the various statistical and dynamical methods for spatial or spatiotemporal downscaling of GPM and TRMM precipitation estimates was conducted, in terms of their advantages and disadvantages, simplicity of application and their suitability at different regional and temporal scales and hydroclimatic conditions. Finally, the adequacy of downscaling remotely sensed precipitation estimates as an effective way to obtain precipitation with sufficient spatial and temporal resolution is discussed and future challenges are highlighted.

I–D Threshold Analysis of Rainfall-Triggered Landslides Based on TRMM Precipitation Data in Wudu, China

This study explored the applicability of TRMM, TRMM nonlinear downscaling, and ANUSPLIN (ANU) interpolation of three different types of precipitation data to define regional-scale rainfall-triggered landslide thresholds. The spatial resolution of TRMM precipitation data was downscaled from 0.25° to 500 m by the downscaling model considering the relationship between humidity, NDVI, and numerous topographic factors and precipitation. The rainfall threshold was calculated using the rainfall intensity–duration threshold model. The calculation showed that TRMM downscaled precipitation data have better detection capability for extreme precipitation events than the other two, the TRMM downscaling threshold was better than the ANU interpolation, and the cumulative effective rainfall of TRMM downscaling was preferred as the macroscopic critical rainfall-triggered landslide threshold for the early warning of the Wudu. The predictive performance of the rainfall threshold of 50% was better than the other two (10% and 90%). When the probability of landslide occurrence was 50%, the TRMM downscaled threshold curve was given by I50=21.03×D−1.004. The authors also analyzed the influence of factors such as topography landform and soil type on the rainfall threshold of landslides in the study area. The rainfall intensity of small undulating mountains was higher than that of medium and large undulating mountains, and the rainfall intensity of landslides peaks at high altitude mountains of 3500–5000 m.

Hydrological drought forecasting and monitoring system development using artificial neural network (ANN) in Ethiopia

The objective of this study is to investigate and perform long–term forecasting of both streamflow and hydrological drought over Ethiopia. Observed streamflow and precipitation data are collected from 17 streamflow stations and 34 rainfall gauge stations to forecast future streamflow and hydrological drought from 2026 to 2099. Streamflow forecasting is performed using an artificial neural network (ANN) in conjunction with python software. Observed precipitation and streamflow data from 1973 to 2014 are used to train and test the ANN model by 70 and 30% ratios, respectively. After training the model, future downscaled precipitation data from regional climate models (RCM) have been used as input data to forecast future streamflow. Three RCM models were used to downscale historical and future climate data. RACMO is found a good downscaling model for all selected stations. The linear scaling bias correction technique results in less than 2% error compared to other alternative techniques. The result indicates that ANN is a good tool to forecast streamflow in areas having a good correlation between precipitation and streamflow such as Abbay, Awash, Baro, Omo Gibe, and Tekeze river basins. But in arid areas for example Genale Dawa, Wabishebele, and Rift Valley basins, the model is not suitable because the input data (precipitation) have high variation than the output variable (streamflow). In such areas, meteorological drought analysis and forecasting are better than hydrological drought analysis. Finally, future hydrological drought is analyzed using forecasted streamflow data as input to the streamflow drought index (SDI). The result indicates that 2028, 2036, 2042, 2044, 2062, and 2063 are the expected extreme drought years in most river basins of Ethiopia in the future. This shows that at least one extreme drought is expected in each decade in the future. Therefore, extensive research in drought analysis and forecasting is needed to develop an effective drought early warning system, and water resource management policy.

A New Downscaling-Calibration Procedure for TRMM Precipitation Data Over Yangtze River Economic Belt Region Based on a Multivariate Adaptive Regression Spline Model

High-resolution precipitation products are essential for accurate hydrological and meteorological applications. To improve the spatial resolution and accuracy of monthly satellite precipitation products, we developed a new downscaling-calibration framework with three key steps: 1) coarse-resolution satellite precipitation data are downscaled to 1-km resolution precipitation data with multivariate adaptive regression spline (MARS) model; 2) residual correction is applied to bridge the difference between the satellite precipitation data and downscaled precipitation data; and 3) the geographical differential analysis (GDA) calibration method is implemented to improve accuracy by merging the residual-corrected data with rain gauge data. In this study, geolocation variables (longitude and latitude), a digital elevation model (DEM) data, daytime and nighttime land surface temperatures, and four remote sensing indices were used to downscale monthly Tropical Rainfall Measuring Mission (TRMM) 3B43 precipitation datasets over the Yangtze River Economic Belt. The downscaled results showed that the MARS model can avoid “boxy artifact” and pixel-level anomalies, which are often found in geographically weighted regression (GWR) and random forest (RF) results. According to the validation, the step of residual correction is not necessary. With GDA calibration, the MARS-based estimated results were more accurate than the results of the other methods (i.e., GWR and RF) and original TRMM products. Therefore, the developed MARS-based downscaling-calibration procedure can improve not only the spatial resolution but also the quality of the TRMM 3B43 products.

Computer Mathematical Statistics Analysis of downscaling data of different vegetation index inversion TRMM

This paper uses enhanced vegetation index (EVI) data, normalized vegetation index (NDVI) data, DEM, aspect data, and TRMM3B43 (V7) data, based on a geographically weighted regression model (GWR), and uses a statistical downscaling method to achieve Central China Downscaling of regional TRMM data from 2010 to 2019. The research results show: (1) TRMM data has good applicability in Central China, and the R2of TRMM data and weather station measured data is above 0.8. (2) Improve the ground resolution from 0.25°×0.25° (approximately 27.5km×27.5km) to 1km×1km while ensuring the same accuracy as the original data. (3) Overall, the accuracy of EVI downscaled precipitation data in Central China is better than that of NDVI downscaled precipitation data.

Reconstructing high-resolution gridded precipitation data using an improved downscaling approach over the high altitude mountain regions of Upper Indus Basin (UIB)

Understanding the basin-scale hydrology and the spatiotemporal distribution of regional precipitation requires high precision, as well as high-resolution precipitation data. We have made an attempt to develop an Integrated Downscaling and Calibration (IDAC) framework to generate high-resolution (1 km × 1 km) gridded precipitation data. Traditionally, GWR (Geographical weighted regression) model has widely been applied to generate high-resolution precipitation data for regional scales. The GWR model generally assumes a spatially varied relationships between precipitation and its associated environmental variables, however, the relationships need to remain constant (fixed) for some variables over space. In this study, a Mixed Geographically Weighted Regression (MGWR) model, capable of dealing with the fixed and spatially varied environmental variables, is proposed to downscale the Original-TRMM precipitation data from a coarse resolution (0.25o × 0.25o) to a high-resolution (1 km × 1 km) for the period of 2000–2018 over the Upper Indus Basin (UIB). Additionally, accuracy of the downscaled precipitation data was further improved by merging it with the recorded data from rain gauge stations (RGS) using two calibration approaches such as Geographical Ratio Analysis (GRA) and Geographical Difference Analysis (GDA). We found MGWR to perform better given its higher R2 and lower RMSE and bias values (R2 = 0.96; RMSE = 56.01 mm, bias = 0.014) in comparison to the GWR model (R2 = 0.95; RMSE = 60.76 mm, bias = 0.094). It was observed that the GDA and GRA calibrated-downscaled precipitation datasets were superior to the Original-TRMM, yet GRA outperformed GDA. Annual precipitation from downscaled and calibrated-downscaled datasets was further temporally downscaled to obtain high-resolution monthly and daily precipitations. The results revealed that the monthly-downscaled precipitation (R2 = 0.82, bias = −0.02 and RMSE = 11.93 mm/month) and the calibrated-downscaled (R2 = 0.89, bias = −0.006 and RMSE = 9.19 mm/month) series outperformed the Original-TRMM (R2 = 0.72, bias = 0.14 and RMSE = 19.8 mm/month) as compared to the RGS observations. The results of daily calibrated-downscaled precipitation (R2 = 0.79, bias = 0.001 and RMSE = 1.7 mm/day) were better than the Original-TRMM (R2 = 0.64, bias = − 0.12 and RMSE = 6.82 mm/day). In general, the proposed IDAC approach is suitable for retrieving high spatial resolution gridded data for annual, monthly, and daily time scales over the UIB with varying climate and complex topography.

Comparison and assessment of spatial downscaling methods for enhancing the accuracy of satellite-based precipitation over Lake Urmia Basin

Estimating precipitation at high spatial-temporal resolution is vital in manifold hydrological, meteorological and water management applications, especially over areas with un-gauged networks and regions where water resources are on the wane. This study aims to evaluate five downscaling methods to determine the accuracy and efficiency of which on generating high-resolution precipitation data at annual and monthly scales. To establish precipitation-Land surface characteristics relationship, environmental factors, including Normalized Difference Vegetation Index (NDVI), Land Surface Temperature (LST) and Digital Elevation Model (DEM), were considered as proxies in the spatial downscaling procedure. The downscaling algorithms, namely support vector machine (SVM), random forest (RF), geographically weighted regression (GWR), multiple linear regression (MLR) and exponential regression (ER), were implemented to downscale the version 7 of TRMM (Tropical Rainfall Measuring Mission) precipitation (3B43 V7 product) over Lake Urmia Basin (LUB) from 0.25° to 1 km spatial resolution. The downscaled precipitation data was validated against observations from meteorological stations. Monthly fractions derived from TRMM 3B43 were used to disaggregate 1 km annual precipitation to 1 km monthly precipitation. Furthermore, the best method was selected for calibration based on Geographical Difference Analysis (GDA) to assess the effectiveness of the calibration as a viable option. The results indicate that SVM not only outperforms the other methods, but also has good agreements with in-situ measurements compared to the original TRMM. The results confirm that inclusion of LST and geographic information along with NDVI can improve the downscaling performance. Downscaling and GDA calibration significantly improve the accuracy of TRMM 3B43 product at both spatial and temporal resolution and should be considered as an essential step in calibration of TRMM precipitation. Calibration at monthly scale yields slightly better results than calibration at annual scale and then disaggregating into monthly maps in terms of accuracy assessment.

Improved flood simulation accuracy by downscaling remotely sensed precipitation data in the Qixing Wetland Watershed

Runoff is an important factor in maintaining the normal operation of wetland ecosystems. An effective runoff simulation can provide an important decision basis for wetland protection and ecological restoration. Aiming to make the spatial resolution of remotely sensed precipitation products match the existing high-precision digital elevation model (DEM), this study used multiple linear regression (MLR) and geographic weighted regression (GWR) methods to build a downscaling correction method for precipitation products in the Qixing Wetland Watershed. The TRMM 3B42V7 data were downscaled from 0.25° × 0.25° to 0.05° × 0.05°. Compared with the downscaling correction of the nearest neighbor method, the GWR correction performed better and could noticeably reduce the average relative deviation between the corrected precipitation series and the measured precipitation series and increase the correlation coefficient of the two series. For the MLR correction, the two indicators (the average relative deviation and the correlation coefficient) improved slightly. The downscaled precipitation data were input into the semidistributed hydrological model. The flood simulation results revealed that the GWR-corrected rainfall data could improve the accuracy of flood simulation better than the MLR.

Optimization and Variants of Quantile-Based Methods for Bias Corrections of Statistically Downscaled Precipitation Data

AbstractNew optimization and variants of quantile-based methods are developed for bias corrections of monthly and daily general circulation model (GCM)-based statistically downscaled precipitation ...

Grid-scale agricultural land and water management: A remote-sensing-based multiobjective approach

This paper developed a remote-sensing-based multiobjective (RSM) approach to formulate sustainable agricultural land and water resources management strategies at a grid scale. To meet the spatial resolution and accuracy need of agricultural management, downscaled precipitation data sets were obtained with the help of global precipitation measurement (GPM) data and other spatial information. Spatial crop water requirement information were obtained via the combination use of the Penman-Monteith method, remote sensing information (MOD16/PET) and virtual water theory. Through integrating these spatial data and considering the impact of different spatial environments on crop growth, a grid-based integer multiobjective programming (GIMP) model was developed to determine best suitable crop planting types at all grids. GIMP can simultaneously consider several conflicting objectives: crop growth suitability, crop spatial water requirements, and ecosystem service value. Further, GIMP results were inputted into a grid-based nonlinear fractional multiobjective programming (GNFMP) model with three objectives: maximize economic benefits, maximize water productivity, and minimize blue water utilization, to optimize irrigation-water allocation. To verify the validity of the proposed approach, a real-world application in the middle reaches of Heihe River Basin, northwest China was conducted. Results show that the proposed method can improve the ecosystem service value by 0.36 × 108 CNY, the economic benefit by 21.85%, the irrigation-water productivity by 25.92%, and reduce blue water utilization rate by 24.32% comparing with status quo.

Spatio-temporal evaluation of various global circulation models in terms of projection of different meteorological drought indices

During the coming years, a dramatic increase is expected in the number of drought events mainly due to climate change. In this study, the spatio-temporal variations of duration and intensity of drought events during two 30-year periods in the future (2040–2069 and 2070–2099) together with the reference period (1971–2000) were investigated based on the impacts of climate change. Three drought indices including the Standardized Precipitation Index, the China Z Index, and the Statistical Z Score were calculated from the dynamically downscaled precipitation data from the outputs of GFDL-ESM2M, HadGEM2-ES, and MPI-ESM-MR general circulation models (GCMs) under RCP4.5 and RCP8.5 scenarios over the Western Black Sea (WBS) and the Euphrates–Tigris (ET) basins in Turkey. The biases in the GCMs' precipitation were corrected using the linear scaling method. Additionally, the Mann–Kendall trend test was adapted to the values of drought indices to detect the trend. The results of the study showed that there is no meaningful variation between the different drought indices and also the outputs of various GCMs in terms of drought properties. On the other hand, based on the drought indices values calculated from the outputs of all GCMs, it was found that drought duration and intensity will increase during the current century. Additionally, it was concluded that by taking the spatial distribution of drought properties over the basins into the account, there is a slight relationship between the geographical elevation and drought properties over the WBS and the ET basins.

Spatial Downscaling of the Tropical Rainfall Measuring Mission Precipitation Using Geographically Weighted Regression Kriging over the Lancang River Basin, China

Satellite-based precipitation products have been widely used to estimate precipitation, especially over regions with sparse rain gauge networks. However, the low spatial resolution of these products has limited their application in localized regions and watersheds. This study investigated a spatial downscaling approach, Geographically Weighted Regression Kriging (GWRK), to downscale the Tropical Rainfall Measuring Mission (TRMM) 3B43 Version 7 over the Lancang River Basin (LRB) for 2001–2015. Downscaling was performed based on the relationships between the TRMM precipitation and the Normalized Difference Vegetation Index (NDVI), the Land Surface Temperature (LST), and the Digital Elevation Model (DEM). Geographical ratio analysis (GRA) was used to calibrate the annual downscaled precipitation data, and the monthly fractions derived from the original TRMM data were used to disaggregate annual downscaled and calibrated precipitation to monthly precipitation at 1 km resolution. The final downscaled precipitation datasets were validated against station-based observed precipitation in 2001–2015. Results showed that: 1) The TRMM 3B43 precipitation was highly accurate with slight overestimation at the basin scale (i.e., CC (correlation coefficient) = 0.91, Bias = 13.3%). Spatially, the accuracies of the upstream and downstream regions were higher than that of the midstream region. 2) The annual downscaled TRMM precipitation data at 1 km spatial resolution obtained by GWRK effectively captured the high spatial variability of precipitation over the LRB. 3) The annual downscaled TRMM precipitation with GRA calibration gave better accuracy compared with the original TRMM dataset. 4) The final downscaled and calibrated precipitation had significantly improved spatial resolution, and agreed well with data from the validated rain gauge stations, i.e., CC = 0.75, RMSE (root mean square error) = 182 mm, MAE (mean absolute error) = 142 mm, and Bias = 0.78% for annual precipitation and CC = 0.95, RMSE = 25 mm, MAE = 16 mm, and Bias = 0.67% for monthly precipitation.

Frequency analyses of extreme precipitation events in Western Black Sea Basin (Turkey) based on climate change projections

The impacts of climate change on extreme precipitation events in the Western Black Sea Basin of Turkey were investigated by using the annual maxima (AM) and peaks over threshold (POT) methods. Daily precipitation data measured between 1971 and 2000 at nine meteorological stations and projected and dynamically downscaled precipitation data from the outputs of GFDL‐ESM2M, HadGEM2‐ES and MPI‐ESM‐MR global circulation models (GCMs) under Representative Concentration Pathway 4.5 (RCP4.5) and RCP8.5 scenarios were used to generate maximum daily rainfall intensity for 2, 5, 10, 20, 50, 100 and 500 year return periods. The outputs of the GCMs were corrected by a modified linear scaling bias correction method to offset the uncertainties of the GCMs before frequency analyses. The datasets were fitted to a generalized extreme value distribution and a generalized Pareto distribution for the AM and POT methods, respectively. Based on a statistical test a reliable goodness‐of‐fit was obtained. The results of frequency analyses of daily storms by using the GCM data, after bias correction, showed that by the end of the current century the magnitude of the storms will increase by 31.29% and 27.00% based on the AM and POT methods respectively under the RCP4.5 scenario and by 43.51% and 31.29% based on the AM and POT methods respectively under the RCP8.5 scenario. The magnitude of the 24 hr rainfall intensity calculated by the AM method was more than that of the POT method by 22% on average. For the whole basin (on average) the mean maximum daily rainfall intensity was found to be 33.45, 46.18, 57.24, 69.16, 87.68, 104.51 and 154.60 mm/day for the respective return periods.

Downscaling rainfall using deep learning long short‐term memory and feedforward neural network

Choosing downscaling techniques is crucial in obtaining accurate and reliable climate change predictions, allowing for detailed impact assessments of climate change at regional and local scales. Traditional statistical methods are likely inefficient in downscaling precipitation data from multiple sources or complex data patterns, so using deep learning, a form of nonlinear models, could be a promising solution. In this study, we proposed to use deep learning models, the so‐called long short‐term memory and feedforward neural network methods, for precipitation downscaling for the Vietnamese Mekong Delta. Model performances were assessed for 2036–2065 period, using original climate projections from five climate models under the Coupled Model Intercomparison Project Phase 5, for two Representative Concentration Pathway scenarios (RCP 4.5 and RCP 8.5). The results exhibited that there were good correlations between the modelled and observed values of the testing and validating periods at two long‐term meteorological stations (Can Tho and Chau Doc). We then analysed extreme indices of precipitation, including the annual maximum wet day frequency (Prcp), 95th percentile of precipitation (P95p), maximum 5‐day consecutive rain (R5d), total number of wet days (Ptot), wet day precipitation (SDII) and annual maximum dry day frequency (Pcdd) to evaluate changes in extreme precipitation events. All the five models under the two scenarios predicted that precipitation would increase in the wet season (June–October) and decrease in the dry season (November–May) in the future compared to the present‐day scenario. On average, the means of multiannual wet season precipitation would increase by 20.4 and 25.4% at Can Tho and Chau Doc, respectively, but in the dry season, these values were projected to decrease by 10 and 5.3%. All the climate extreme indices would increase in the period of 2036–2065 in comparison to the baseline. Overall, the developed downscaling models can successfully reproduce historical rainfall patterns and downscale projected precipitation data.

Downscaling Satellite Precipitation Estimates With Multiple Linear Regression, Artificial Neural Networks, and Spline Interpolation Techniques

AbstractSatellite precipitation estimates (SPEs) have been widely used in various applications. However, when applied to small basins and regions, the spatial resolution of SPEs is too coarse. In this study, we present three downscaling algorithms based upon the relationships between SPEs and cloud optical and microphysical properties in northeast Austria. Different downscaling techniques, namely, multiple linear regression, artificial neural networks, and spline interpolation, were adopted for the downscaling of Integrated Multi‐satellitE Retrievals for GPM (IMERG) precipitation data. In this respect, linear and nonlinear relationship among IMERG data and different cloud variables, such as cloud effective radius, cloud optical thickness, and cloud water path, was evaluated. Downscaled SPEs, as well as the original IMERG product, were subsequently validated using 54 rain gauges at a daily timescale. According to the results, all downscaled products were more accurate than the original IMERG data. Furthermore, all downscaling techniques captured the spatial patterns of precipitation reasonably well with more detailed information when compared with the original IMERG precipitation. However, the spline interpolation slightly outperformed the other techniques, followed by multiple linear regression and artificial neural network, respectively. Moreover, the proposed methods, which consistently showed increased correlation (e.g., from 0.30 to 0.56 for spline interpolation) and reduced mean absolute error and root‐mean‐square error (e.g., from 10.14 to 6.55 mm and 13.5 to 8.76 mm, respectively) for average of all events, can more accurately produce downscaled precipitation data.

Evaluation of physical parameterizations for atmospheric river induced precipitation and application to long-term reconstruction based on three reanalysis datasets in Western Oregon

Dynamically downscaled precipitation is often used for evaluating sub-daily precipitation behavior on a watershed-scale and for the input to hydrological modeling because of its increasing accuracy and spatiotemporal resolution. Despite these advantages, physical parameterizations in regional models and systematic biases due to the dataset used for boundary conditions greatly influence the quality of downscaled precipitation data. The present paper aims to evaluate the performance and the sensitivities of physical parameterizations of the Weather Research and Forecasting (WRF) model to simulate extreme precipitation associated with atmospheric rivers (ARs) over the Willamette watershed in Oregon. Also investigated was whether the optimized WRF configuration for extreme events can be used for long-term reconstruction using different boundary condition datasets. Three reanalysis datasets, the Twentieth Century Reanalysis version 2c (20CRv2c), the European Center for Medium-Range Weather Forecasts (ECMWF) twentieth century reanalysis (ERA20C), and the Climate Forecast System Reanalysis (CFSR), which have different spatial resolutions and dataset periods, were used to simulate precipitation at 4 km resolution. Sensitivity analyses showed that AR precipitation is most sensitive to the microphysics parameterization. Among 13 microphysics schemes investigated, the Goddard and the Stony-Brook University schemes performed the best regardless of the choice of reanalysis. Reconstructed historical precipitation with the optimized configuration showed better accuracies during the wet season than the dry season. With respect to simulations with CFSR, it was found that the optimized configuration for AR precipitation can be used for long-term reconstruction with small biases. However, systematic biases in the reanalysis datasets may still lead to uncertainties in downscaling precipitation in a different season with a single configuration.

Mapping Fine Spatial Resolution Precipitation from TRMM Precipitation Datasets Using an Ensemble Learning Method and MODIS Optical Products in China

Precipitation data are important for the fields of hydrology and meteorology, and are fundamental for ecosystem monitoring and climate change research. Satellite-based precipitation products are already able to provide high temporal resolution precipitation information at a global level. However, the coarse spatial resolution has restricted their use in regional level studies. In this study, monthly fine spatial resolution land precipitation data in China was obtained by downscaling the TRMM 3B43 V7 monthly precipitation products. The downscaling model was constructed based on the ensemble learning method called random forest (RF). In addition to the RF model, the classification and regression tree (CART) model was also used to downscale the precipitation data for the purpose of comparison. The results were validated with in situ measurements. Results showed that the RF model outperformed the CART model. The downscaled precipitation data were strongly correlated with the in situ measurements. The downscaling method was applied to mapping fine spatial resolution precipitation over all of China, and is valuable for developing high spatial resolution precipitation products for studies on hydrology, meteorology, and climate science.