Precipitation Estimation From Remotely Sensed Information Using Artificial Neural Networks-Cloud Classification System Research Articles

Performance of the PERSIANN Family of Products over the Mekong River Basin and Their Application for the Analysis of Trends in Extreme Precipitation Indices

Near-real-time satellite precipitation estimation is indispensable in areas where ground-based measurements are not available. In this study, an evaluation of two near-real-time products from the Center for Hydrometeorology and Remote Sensing at the University of California, Irvine—PERSIANN-CCS (Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks—Cloud Classification System) and PDIR-Now (PERSIANN-Dynamic Infrared Rain Rate near-real-time)—were compared to each other and evaluated against IMERG Final (Integrated Multi-satellite Retrievals for Global Precipitation Measurement—Final Run) from 2015 to 2020 over the Mekong River Basin and Delta (MRB) using a spatial resolution of 0.1∘ by 0.1∘ and at a daily scale. PERSIANN-CDR (PERSIANN-Climate Data Record) was also included in the evaluation but was not compared against the real-time products. In this evaluation, PDIR-Now exhibited a superior performance to that of PERSIANN-CCS, and the performance of PERSIANN-CDR was deemed satisfactory. The second part of the study entailed performing a Mann–Kendall trend test of extreme precipitation indices using 38 years of PERSIANN-CDR data over the MRB. This annual trend analysis showed that extreme precipitation over the 95th and 99th percentiles has decreased over the Upper Mekong River Basin, and the consecutive number of wet days has increased over the Lower Mekong River Basin.

Performance evaluation of multiple satellite rainfall data sets in central highlands of Abbay Basin, Ethiopia

ABSTRACT This study evaluates four satellite-based precipitation datasets with gauged rainfall observations at a daily and wet season time scales. Satellite Precipitation Estimators from Climate Hazards Group Infrared Precipitation Stations (CHIRPSv2), the Climate prediction center (CPC) morphing technique (CMORPH), the Integrated Multi-satellite Retrieval for GPM (IMERG-06) and Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks Cloud Classification System (PERSSIANN-CCS) were used. Categorical and continuous statistical techniques were applied to evaluate the detection and estimation ability of SPEs. Among the SPEs, CMORPH was superior, 0.90 and followed by IMERG-06, 0.82 in the Probability of detection (POD), Critical success index (CSI), 0.74 and 0.70 and Accuracy, 0.75 and 0.72 at wet season, respectively. The result of bias ratio showed underestimation for all SPEs at wet season and daily time scale. The comparison of SPEs in terms of RMSE indicated that IMERG-06 was better, 75.9 followed by CMORPH, 86.0. In terms of Relative bias (RBias), IMERG-06 was the second next to CHIRPSv2. On the other hand, IMERG-06 had better RMSE at JJAS and daily time scale. Overall, the findings of this evaluation study showed the capacity of IMERG-06 product could be reliable rainfall data sources for climate and hydrological analysis of central highlands of Abbay Basin.

Precipitation Estimation Based on Infrared Data with a Spherical Convolutional Neural Network

Abstract Precipitation is a vital process in the water cycle. Accurate estimation of the precipitation rate underpins the success of hydrological simulations, flood predictions, and water resource management. Satellite infrared (IR) data, with high temporal resolution and wide coverages, have been commonly used in precipitation inversion. However, existing IR-based precipitation retrieval algorithms suffer from various problems such as overestimation in dry regions, poor performance in extreme rainfall events, and reliance on an empirical cloud-top brightness–rain rate relationship. To resolve these problems, we construct a deep learning model using a spherical convolutional neural network to properly represent Earth’s spherical surface. With data input directly from IR bands 3, 4, and 6 of the operational Geostationary Operational Environmental Satellite (GOES), our new model of Precipitation Estimation based on IR data with Spherical Convolutional Neural Network (PEISCNN) was first trained and tested with a 3-month-long dataset, and then validated in a 2-yr period. Compared to the commonly used IR-based precipitation product PERSIANN-CCS (Precipitation Estimation from Remotely Sensed Information Using Artificial Neural Networks–Cloud Classification System), PEISCNN showed significant improvement in the metrics of POD, CSI, RMSE, and CC, especially in the dry region and for extreme rainfall events. Decomposed with the four-component error decomposition (4CED) method, the overestimation of PEISCNN was averaged 47.66% lower than the CCS at the hourly scale. The PEISCNN model may provide a promising way to produce an improved IR-based precipitation product to benefit a wide range of hydrological applications. Significance Statement An IR-based precipitation algorithm is irreplaceable in satellite precipitation inversion, since an IR sensor can provide observations of high frequency, fine temporal resolution, and wide coverage. Considering the spherical nature of Earth’s surface which has been overlooked in previous IR-based precipitation retrieval algorithms, we proposed a new deep learning model PEISCNN, which can address the problems that exist in IR-based precipitation estimations such as overestimation in dry regions, deficiency in extreme rainfall events, and reliance on the empirical cloud-top brightness–rain rate relationship. PEISCNN provides a new insight to improve the accuracy of the satellite IR-based or multisensor-based precipitation estimation, and it has great potential to benefit a range of related hydrological research, applications in water resource management, and flood predictions.

Hypothetical Ground Radar-Like Rain Rate Generation of Geostationary Weather Satellite Using Data-to-Data Translation

This study proposes a deep-learning-based data-to-data (D2D) translation framework to simulate a radar-like retrieval of rainfall rates using the advantages of spatial coverage and temporal resolution of geostationary (GEO) satellite observation. The D2D method comprises normalization and denormalization in pre- and post-processing and an adversarial learning structure for an inter-domain conversion between physical values of data such as albedo and brightness temperature (BT) unlike the image-to-image translation using digital number values in image data. The GEO-KOMPSAT-2A (GK2A) and radar hybrid surface rainfall (HSR) datasets over the Korean Peninsula from September 2019 to September 2021 were used as the source and target domains for training and testing the D2D model. The constructed D2D model for ground radar-like rainfall generation was validated using the ground radar-observed rainfall data and compared to the GK2A rainfall rate (RR), Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Cloud Classification System (PERSIANN-CCS), and Integrated Multi-satellite Retrievals for Global Precipitation Measurement (IMERG) rainfall products. The D2D model exhibited excellent performance for various rain types in the study area compared to the GK2A RR, PERSIANN-CCS, and IMERG data. Consequently, the D2D model can provide valuable and accurate radar-like rainfall intensity and distribution data with a high temporal resolution and complementary rainfall information over lands and oceans without radar observation.

A framework for comparing two rainfields based on spatial structure: A case of radar against selected satellite precipitation products over southeast Queensland, Australia

• Rain gauge networks are not adequate to capture the complex spatial structure. • Radar and satellite precipitation products capture the spatial structure of rainfall. • Fine spatial scale rainfall data can be derived from satellite precipitation products. • SSI and correlogram are good tools for comparing two rainfields. • Some freely available satellite precipitation products are as good as radar rainfall. Rain gauge networks are not adequate to capture the complex spatial structure of rainfall for meaningful urban flood studies. While radar and satellite precipitation products (SPPs) are being developed to alleviate the limitations of rain gauge networks, they need to be evaluated for regional applications. This study compared radar against 3 SPPs of IMERG (Integrated Multi-satellite Retrievals for Global Precipitation Measurement), MSWEP (Multi-Source Weighted-Ensemble Precipitation) and PERSIANN-CCS (Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks Cloud Classification System) that have sub-daily records for southeast Queensland, Australia, based on the spatial structure identified through the structural similarity index (SSI) and two-dimensional correlogram at the daily timescale. The SPPs were downscaled to 1 km × 1 km grid resolution to conform with the radar dataset using inverse distance weighting and bilinear interpolation. All three SPPs, that are freely available, reflected very well the similarity (>90 %) in the spatial pattern with that of the radar, although IMERG was marginally better. However, MSWEP appeared to be the closest to the gauge dataset, better than even the expensive radar dataset, stemming from its much better similarity in the mean similarity statistic. Nevertheless, there could be some atmospheric climatic factors that may cause one to select one SPP over the others for a given wet day, and this needs to be explored for the possibility of selecting different SPPs for different wet days.

Reliability of Gridded Precipitation Products for Water Management Studies: The Case of the Ankavia River Basin in Madagascar

Hydrological modeling for water management in large watersheds requires accurate spatially-distributed rainfall time series. In case of low coverage density of ground-based measurements, gridded precipitation products (GPPs) from merged satellite-/gauge-/model-based rainfall products constitute an attractive alternative. The quality of which must, nevertheless, be verified. The objective of this study was to evaluate, at different time scales, the reliability of 6 GPPs against a 2-year record from a network of 14 rainfall gauges located in the Ankavia catchment (Madagascar). The GPPs considered in this study are the African Rainfall Estimate Climatology (ARC2), the Climate Hazards Group Infrared Precipitation with Station data (CHIRPS), the European Centre Medium-Range Weather Forecasts ECMWF Reanalysis on global land surface (ERA5-Land), the Integrated Multi-satellitE Retrievals for Global Precipitation Measurement V06 Final (IMERG), the Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks Cloud Classification System (PERSIANN-CCS), and the African Rainfall Estimation (RFEv2) products. The results suggest that IMERG (R2 = 0.63, slope of linear regression a = 0.96, root mean square error RMSE = 12 mm/day, mean absolute error MAE = 5.5 mm/day) outperforms other GPPs at the daily scale, followed by RFEv2 (R2 = 0.41, a = 0.94, RMSE = 15 mm/day, MAE = 6 mm/day) and ARC2 (R2 = 0.30, a = 0.88, RMSE = 16 mm/day, MAE = 6.7 mm/day). All GPPs, with the exception of the ERA5, overestimate the ‘no rain’ class (0–0.2 mm/day). ARC2, IMERG, PERSIANN, and RFEv2 all underestimate rainfall occurrence in the 0.2–150 mm/day rainfall range, whilst CHIRPS and ERA5 overestimate it. Only CHIRPS and PERSIANN could estimate extreme rainfall (>150 mm/day) satisfactorily. According to the Critical Success Index (CSI) categorical statistical measure, IMERG performs quite well in detecting rain events in the range of 2–100 mm/day, whereas PERSIANN outperforms IMERG for rain events larger than 150 mm/day. Because it performs best at daily scale, only IMERG was evaluated for time scales other than daily. At the yearly and monthly time scales, the performance is good with R2 = 0.97 and 0.87, respectively. At the event time scale, the probability distribution function PDF of rain gauge values and IMERG data show good agreement. However, at an hourly time scale, the correlation between ground-based measurements and IMERG data becomes poor (R2 = 0.20). Overall, the IMERG product can be regarded as the most reliable gridded precipitation source at monthly, daily, and event time scales for hydrological applications in the study area, but the poor agreement at hourly time scale and the inability to detect extreme rainfall >100 mm/day may, nevertheless, restrict its use.

Can GPM IMERG Capture Extreme Precipitation in North China Plain?

Extreme precipitation events (EPE) often cause catastrophic floods accompanied by serious economic losses and casualties. The latest version (V06) of the Integrated Multi-satellite Retrievals for Global Precipitation Measurement (GPM IMERG) provides global satellite precipitation data from 2000 at a higher spatiotemporal resolution with improved quality. It is scientifically and practically important to assess the accuracy of the IMERG V06 in capturing extreme precipitation. This study evaluates the two widely used products of IMERG during 2000–2018, i.e., IMERG late run (IMERG-L) and IMERG final run (IMERG-F), in the densely populated and flood-prone North China Plain. The accuracy of the IMERG V06 is evaluated with ground measurements from rain gauge stations at multiple scales (hourly, daily, and seasonally). A novel target tracking method is introduced to extract three-dimensional (3D) extreme precipitation events, and the near-real-time uncalibrated PERSIANN-CCS (Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks Cloud Classification System) and GSMAP (Global Satellite Mapping of Precipitation) satellite data are added to further evaluate IMERG’s performance during extreme precipitation. Finally, for flash flood events induced by extreme rainfall in the Hebei Province from 15 to 23 July 2016, the accuracy of capturing the event with IMERG-F and IMERG-L was verified. Results reveal that IMERG-F is better than IMERG-L at all investigated scales (hourly, daily, and seasonally), but the difference between the two products is less at higher time resolutions. Both products manifest decreased performance when capturing 3D extreme precipitation events, and comparatively, IMERG-F performs better than IMERG-L. IMERG-F exhibits a distinct discontinuity in extreme precipitation thresholds between land and ocean, which is a limitation of IMERG-F not documented in previous studies. Moreover, IMERG-L and IMERG-F are comparable at an hourly scale for some metrics, which is beyond the expectation that IMERG-F is notably better than IMERG-L. This study provides a scientific basis for the performance of satellite precipitation products and contributes to guiding users when applying global precipitation products.

Simulation of Karst Floods with a Hydrological Model Improved by Meteorological Model Coupling

Abstract Karst basins are prone to rapid flooding because of their geomorphic complexity and exposed karst landforms with low infiltration rates. Accordingly, simulating and forecasting floods in karst regions can provide important technical support for local flood control. The study area, the Liujiang karst river basin, is the most well-developed karst areas in South China, and its many mountainous areas lack rainfall gauges, limiting the availability of precipitation information. Quantitative precipitation forecast (QPF) from the Weather Research and Forecasting (WRF) Model and quantitative precipitation estimation (QPE) from the Precipitation Estimation from Remotely Sensed Information Using Artificial Neural Networks–Cloud Classification System (PERSIANN-CCS) can offer reliable precipitation estimates. Here, the distributed Karst–Liuxihe (KL) model was successfully developed from the terrestrial Liuxihe model, as reflected in improvements to its underground structure and confluence algorithm. Compared with other karst distributed models, the KL model has a relatively simple structure and small modeling data requirements, which are advantageous for flood prediction in karst areas lacking hydrogeological data. Our flood process simulation results suggested that the KL model agrees well with observations and outperforms the Liuxihe model. The average Nash coefficient, correlation coefficient, and water balance coefficient increased by 0.24, 0.19, and 0.20, respectively, and the average flood process error, flood peak error, and peak time error decreased by 13%, 11%, and 2 h, respectively. Coupling the WRF Model and PERSIANN-CCS with the KL model yielded a good performance in karst flood simulation and prediction. Notably, coupling the WRF and KL models effectively predicted the karst flood processes and provided flood prediction results with a lead time of 96 h, which is important for flood warning and control. Significance Statement The WRF Model and PERSIANN-CCS can provide precipitation data for mountainous karst areas lacking rainfall gauges, and their rainfall results are forecasted effectively to reduce the uncertainty of input precipitation data. Then, the PERSIANN-CCS QPEs and WRF QPF are coupled with the improved KL model for karst flood simulation and prediction. This coupled model worked well in karst basins.

Infrared precipitation estimation using convolutional neural network for FengYun satellites

Infrared (IR) is an important data source for satellite quantitative precipitation estimation, and has been widely applied in the fields of meteorology, hydrology, and agriculture. In the past decades, a series of IR retrieval algorithms have been developed to support the production of IR-based and IR-microwave merged precipitation products. Recently, deep learning techniques such as the convolutional neural network (CNN) show great potential in obtaining IR precipitation estimates with higher accuracy than traditional retrieval algorithms. In this study, we present an upgraded version of IR Precipitation Estimation using CNN (IPEC), i.e., IPEC version 2 (IPEC-V2), with a new end-to-end manner. IPEC-V2 is used to generate a precipitation dataset based on IR data from China’s Fengyun (FY) geostationary satellites (FY-2F, 2G, and 4A). To overcome the difficulty of model training in regions with sparse observations, IPEC-V2 models are firstly pre-trained over the data-rich Continental US (CONUS) using IR data from the Geostationary Operational Environment Satellite (GOES). The models are then transferred to China through re-training with multi-band IR signals from FY satellites. Finally, a long-term record of high-resolution FY IR precipitation estimates is produced during the period from November 9, 2012 to February 28, 2021, which is named as IPEC-FY. IPEC-FY shows better performance than the baseline Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Cloud Classification System (PERSIANN-CCS) with 76.2% gain in Pearson’s correlation coefficient and 18.4% gain in root mean squared error in China. This study shows that transfer learning is an effective way to build CNN models in regions without enough observations, and the high-quality IPEC-FY can act as an alternative precipitation dataset for research in China.

Appraising the Potential of Using Satellite‐Based Rainfall Estimates for Evaluating Extreme Precipitation: A Case Study of August 2014 Event Across the West Rapti River Basin, Nepal

AbstractHeavy precipitation events are recurrently occurring in Nepal, affecting lives and properties every year, especially in the summer monsoon season (i.e., June‐September). We investigated an extreme (heavy) precipitation event of August 2014 over the West Rapti River (WRR) Basin, Nepal. First, we forced a rainfall‐runoff model with ground‐based (gauge) hourly rainfall data of nine stations. Second, we validated against hourly water level at an outlet of the WRR Basin. This study then evaluated the performance of different satellite‐based rainfall estimates (SREs) in capturing an extreme precipitation event. We considered the use of half‐hourly data of Integrated Multi‐satellite Retrievals for GPM (IMERG) (Early, Late, and Final versions), spatial resolution (10 km), and hourly data of Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks (PERSIANN), spatial resolution (25 km), and Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks‐Cloud Classification System (PERSIANN‐CCS), spatial resolution (4 km). Also, we used 3 h data of Tropical Multi‐satellite Precipitation Analysis (TMPA) product real‐time (3B42RT), spatial resolution (25 km). In general, we find that all selected SREs depicted a similar pattern of extreme precipitation as shown by the gauge data on a daily scale. However, we find these products could not replicate precisely on a sub‐daily scale. Overall, IMERG and TMPA showed a better performance than PERSIANN and PERSIANN‐CCS. Finally, we corrected poor‐performed SREs with respect to gauge data and also filled data gaps of gauge rainfall using the information of good‐performed SREs. Our study reveals that there is a great challenge in local flood simulation employing SREs at high‐temporal resolution in Nepal.

Application of remote sensing precipitation data and the CONNECT algorithm to investigate spatiotemporal variations of heavy precipitation: Case study of major floods across Iran (Spring 2019)

In recent years, the number of floods following unprecedented rainfall events have increased in Iran during early spring (March 21st to April 20th, referred to in Iran as the month of “Farvadin”). While numerous studies have addressed changes in climate extremes and precipitation trends at different temporal scales from daily to annual across the country, analyses of short-duration and heavy precipitation, especially during recent years, are rarely considered. Furthermore, most studies investigate the variations in extremes and total precipitation using a limited number of synoptic weather stations across Iran. This study assesses the variations in heavy precipitation (precipitation with intensities greater than or equal to 3 mm/3 h) at 0.04° spatial and 3-hourly temporal resolution during the month of Farvardin. In addition, the effect of atmospheric river conditions over Iran and their possible link to intensifying heavy precipitation is explored. For this purpose, the CONNected-objECT (CONNECT) algorithm is applied on a precipitation dataset, Precipitation Estimation from Remotely Sensed Information Using Artificial Neural Networks-Cloud Classification System (PERSIANN-CCS), and an Integrated Water Vapor Transport (IVT) dataset from the NASA Modern-Era Retrospective Analysis for Research and Applications Version-2 (MERRA-2). The results suggest that the increase in the number of floods in recent years is related to the increase in the intensity and volume of heavy precipitation events, although the frequency and duration of heavy precipitation events have not changed significantly. Furthermore, the results show that atmospheric river conditions over the country are present during the same window as each year’s most extreme events. It is found that 8 out of 13 of the largest ARs over Iran come from moisture plumes with pathways over the African and Red Sea.

CNN-based near-real-time precipitation estimation from Fengyun-2 satellite over Xinjiang, China

Real-time precipitation estimation from satellite observations is still a challenging issue. In this paper, we propose an effective and efficient model based on convolutional neural network (CNN) to estimate precipitation over Xinjiang, the driest region in China, from Fengyun-2 satellite data. Our network mainly consists of two modules: the precipitation identification module and the precipitation estimation module. The first module aims at identifying the given region as precipitation or non-precipitation, while the second one focuses on estimating the specific precipitation amount of the identified region. In addition, considering that topography generally have effects on the quantitative precipitation estimation (QPE) task, we therefore incorporate them into the second module. To evaluate the effectiveness of our proposed model over the national stations, we compare it with not only the Integrated Multi-satellite Retrievals for the Global Precipitation Measurement (GPM) mission (IMERG), but also another two near-real-time products named Global Satellite Mapping of Precipitation Near-Real-Time product (GSMaP_NRT) and Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Cloud Classification System (PERSIANN-CCS). Besides, two precipitation events are selected to validate the generalization ability of our proposed model in Xinjiang, and the verification experiment is conducted upon the precipitation product IMERG Final run. Experimental results show that, in comparison with the rain gauges records, the precipitation identification accuracy and correlation coefficient of our proposed model are 0.93 and 0.21, while the precipitation estimation RMSE and MSE of our proposed model are 0.25 and 0.038 mm/h, respectively. For the two precipitation events, our proposed model demonstrates high consistency with IMERG Final run in precipitation identification, but it shows a tendency to underestimate the heavy rainfall estimated by IMERG Final run. More importantly, the implementation time of our proposed model on the hourly testing dataset is only 5.5 s. Therefore, our proposed model can achieve near-real-time results, which is valuable for precipitation nowcasting applications.

Bias Correction of Satellite-Based Precipitation Estimations Using Quantile Mapping Approach in Different Climate Regions of Iran

High-resolution real-time satellite-based precipitation estimation datasets can play a more essential role in flood forecasting and risk analysis of infrastructures. This is particularly true for extended deserts or mountainous areas with sparse rain gauges like Iran. However, there are discrepancies between these satellite-based estimations and ground measurements, and it is necessary to apply adjustment methods to reduce systematic bias in these products. In this study, we apply a quantile mapping method with gauge information to reduce the systematic error of the Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Cloud Classification System (PERSIANN-CCS). Due to the availability and quality of the ground-based measurements, we divide Iran into seven climate regions to increase the sample size for generating cumulative probability distributions within each region. The cumulative distribution functions (CDFs) are then employed with a quantile mapping 0.6° × 0.6° filter to adjust the values of PERSIANN-CCS. We use eight years (2009–2016) of historical data to calibrate our method, generating nonparametric cumulative distribution functions of ground-based measurements and satellite estimations for each climate region, as well as two years (2017–2018) of additional data to validate our approach. The results show that the bias correction approach improves PERSIANN-CCS data at aggregated to monthly, seasonal and annual scales for both the calibration and validation periods. The areal average of the annual bias and annual root mean square errors are reduced by 98% and 56% during the calibration and validation periods, respectively. Furthermore, the averages of the bias and root mean square error of the monthly time series decrease by 96% and 26% during the calibration and validation periods, respectively. There are some limitations in bias correction in the Southern region of the Caspian Sea because of shortcomings of the satellite-based products in recognizing orographic clouds.

Post and near real-time satellite precipitation products skill over Karkheh River Basin in Iran

ABSTRACT Due to high spatial and temporal resolution and near real-time accessibility of satellite precipitation data, the necessity of using these data in the hydrological application seems to be more pressing than ever. In this study, the skill of six post real-time (Climate Hazards Group Infrared Precipitation with Station data (CHIRPS); CPC MORPHing technique (CMORPH); Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks (PERSIANN); PERSIANN Climate Data Record (PERSIANN-CDR); precipitation produced from the inversion of the satellite soil moisture (SM) observations derived from the European Space Agency (ESA) Climate Change Initiative (SM2RAIN); Tropical Rainfall Measuring Mission (TRMM 3B42-V7)) and two near real-time (PERSIANN Cloud Classification System (PERSIANN-CCS); TRMM real-time (TRMM 3B42-RT)) satellite daily precipitation products are evaluated by comparing with 28 rain gauges in Karkheh River Basin, located in the semi-arid region of Iran. The evaluation is performed for two types of quantiles (lower quantile (< Q10 and < Q25) and upper quantile (> Q50, > Q75, and > Q95)) and rainy seasons using categorical and quantitative metrics for the period March 2003 to December 2014. The spatial analysis indicated that there is not remarkable variation in the skill of satellite precipitation products across the study area. Results showed that the satellite precipitation estimates are more accurate in lower than upper quantile. The seasonal analysis presented that the skill of satellite precipitation products for fall and spring is slightly higher than winter. For post real-time satellite, in terms of POD (VHI), PERSIANN-CDR in spring (winter and spring), SM2RAIN in winter and spring (fall) shows the best skill, and according to FAR and CSI, CMORPH is the best in all seasons. In addition, VHI and POD of PERSIANN-CCS have better skill than 3B42-RT for near real-time satellite for all seasons. Generally, PERSIANN-CCS (PERSIANN-CDR and SM2RAIN) shows the best skill for near (post) real-time satellite precipitation estimations when whole data are included in the analysis.

Infrared Precipitation Estimation Using Convolutional Neural Network

Infrared (IR) information is fundamental to global precipitation estimation. Although researchers have developed numerous IR-based retrieval algorithms, there is still plenty of scope for promoting their accuracy. This article develops a novel deep learning-based algorithm entitled infrared precipitation estimation using a convolutional neural network (IPEC). Based on the five-channel IR data, the IPEC first identifies the precipitation occurrence and then estimates the precipitation rates at hourly and $0.04^{\circ } \times 0.04^{\circ }$ resolutions. The performance of the IPEC is validated using the Stage-IV radar–gauge-combined data and compared to the Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Cloud Classification System (PERSIANN-CCS) in three subregions over the continental United States (CONUS). The results show that the five-channel input is more efficient in precipitation estimation than the commonly used one-channel input. The IPEC estimates based on the five-channel input show better statistical performance than the PERSIANN-CCS with 34.9% gain in Pearson’s correlation coefficient (CC), 38.0% gain in relative bias (BIAS), and 45.2% gain in mean squared error (MSE) during the testing period from June to August 2014 over the central CONUS. Furthermore, the optimized IPEC model is applied in totally independent periods and regions, and still achieves significantly better performance than the PERSIANN-CCS, indicating that the IPEC has a stronger generalization capability. On the whole, this article proves the effectiveness of the convolutional neural network (CNN) combined with the physical multichannel inputs in IR precipitation retrieval. This end-to-end deep learning algorithm shows the potential for serving as an operational technique that can be applied globally and provides a new perspective for the future development of satellite precipitation retrievals.

Deep Neural Network Cloud-Type Classification (DeepCTC) Model and Its Application in Evaluating PERSIANN-CCS

Satellite remote sensing plays a pivotal role in characterizing hydrometeorological components including cloud types and their associated precipitation. The Cloud Profiling Radar (CPR) on the Polar Orbiting CloudSat satellite has provided a unique dataset to characterize cloud types. However, data from this nadir-looking radar offers limited capability for estimating precipitation because of the narrow satellite swath coverage and low temporal frequency. We use these high-quality observations to build a Deep Neural Network Cloud-Type Classification (DeepCTC) model to estimate cloud types from multispectral data from the Advanced Baseline Imager (ABI) onboard the GOES-16 platform. The DeepCTC model is trained and tested using coincident data from both CloudSat and ABI over the CONUS region. Evaluations of DeepCTC indicate that the model performs well for a variety of cloud types including Altostratus, Altocumulus, Cumulus, Nimbostratus, Deep Convective and High clouds. However, capturing low-level clouds remains a challenge for the model. Results from simulated GOES-16 ABI imageries of the Hurricane Harvey event show a large-scale perspective of the rapid and consistent cloud-type monitoring is possible using the DeepCTC model. Additionally, assessments using half-hourly Multi-Radar/Multi-Sensor (MRMS) precipitation rate data (for Hurricane Harvey as a case study) show the ability of DeepCTC in identifying rainy clouds, including Deep Convective and Nimbostratus and their precipitation potential. We also use DeepCTC to evaluate the performance of the Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Cloud Classification System (PERSIANN-CCS) product over different cloud types with respect to MRMS referenced at a half-hourly time scale for July 2018. Our analysis suggests that DeepCTC provides supplementary insights into the variability of cloud types to diagnose the weakness and strength of near real-time GEO-based precipitation retrievals. With additional training and testing, we believe DeepCTC has the potential to augment the widely used PERSIANN-CCS algorithm for estimating precipitation.

Application of Satellite Rainfall Products for Flood Inundation Modelling in Kelantan River Basin, Malaysia

The advent of satellite rainfall products can provide a solution to the scarcity of observed rainfall data. The present study aims to evaluate the performance of high spatial-temporal resolution satellite rainfall products (SRPs) and rain gauge data in hydrological modelling and flood inundation mapping. Four SRPs, Integrated Multi-satellitE Retrievals for Global Precipitation Measurement (GPM) - Early, - Late (IMERG-E, IMERG-L), Global Satellite Mapping of Precipitation-Near Real Time (GSMaP-NRT), and Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks- Cloud Classification System (PERSIANN-CCS) and rain gauge data were used as the primary input to a hydrological model, Rainfall-Runoff-Inundation (RRI) and the simulated flood level and runoff were compared with the observed data using statistical metrics. GSMaP showed the best performance in simulating hourly runoff with the lowest relative bias (RB) and the highest Nash-Sutcliffe efficiency (NSE) of 4.9% and 0.79, respectively. Meanwhile, the rain gauge data was able to produce runoff with −12.2% and 0.71 for RB and NSE, respectively. The other three SRPs showed acceptable results in daily discharge simulation (NSE value between 0.42 and 0.49, and RB value between −23.3% and −31.2%). The generated flood map also agreed with the published information. In general, the SRPs, particularly the GSMaP, showed their ability to support rapid flood forecasting required for early warning of floods.

Effective Cloud Detection and Segmentation Using a Gradient-Based Algorithm for Satellite Imagery: Application to Improve PERSIANN-CCS

Abstract The effective identification of clouds and monitoring of their evolution are important toward more accurate quantitative precipitation estimation and forecast. In this study, a new gradient-based cloud-image segmentation algorithm is developed using image processing techniques. This method integrates morphological image gradient magnitudes to separate cloud systems and patches boundaries. A varying scale kernel is implemented to reduce the sensitivity of image segmentation to noise and to capture objects with various finenesses of the edges in remote sensing images. The proposed method is flexible and extendable from single to multispectral imagery. Case studies were carried out to validate the algorithm by applying the proposed segmentation algorithm to synthetic radiances for channels of the Geostationary Operational Environmental Satellite (GOES-16) simulated by a high-resolution weather prediction model. The proposed method compares favorably with the existing cloud-patch-based segmentation technique implemented in the Precipitation Estimation from Remotely Sensed Information Using Artificial Neural Networks–Cloud Classification System (PERSIANN-CCS) rainfall retrieval algorithm. Evaluation of event-based images indicates that the proposed algorithm has potentials comparing to the conventional segmentation technique used in PERSIANN-CCS to improve rain detection and estimation skills with an accuracy rate of up to 98% in identifying cloud regions.

Bias adjustment of satellite-based precipitation estimation using artificial neural networks-cloud classification system over Saudi Arabia

Precipitation is a key input variable for hydrological and climate studies. Rain gauges can provide reliable precipitation measurements at a point of observations. However, the uncertainty of rain measurements increases when a rain gauge network is sparse. Satellite-based precipitation estimations SPEs appear to be an alternative source of measurements for regions with limited rain gauges. However, the systematic bias from satellite precipitation estimation should be estimated and adjusted. In this study, a method of removing the bias from the precipitation estimation from remotely sensed information using artificial neural networks-cloud classification system (PERSIANN-CCS) over a region where the rain gauge is sparse is investigated. The method consists of monthly empirical quantile mapping of gauge and satellite measurements over several climate zones as well as inverse-weighted distance for the interpolation of gauge measurements. Seven years (2010–2016) of daily precipitation estimation from PERSIANN-CCS was used to test and adjust the bias of estimation over Saudi Arabia. The first 6 years (2010–2015) are used for calibration, while 1 year (2016) is used for validation. The results show that the mean yearly bias is reduced by 90%, and the yearly root mean square error is reduced by 68% during the validation year. The experimental results confirm that the proposed method can effectively adjust the bias of satellite-based precipitation estimations.

Probabilistic precipitation rate estimates with space‐based infrared sensors

The uncertainty structure of satellite‐based passive infrared quantitative precipitation estimation (QPE) is largely unknown at fine spatio‐temporal scales, and requires more than just one deterministic “best estimate” to adequately cope with the intermittent, highly skewed distribution that characterizes precipitation. An investigation of this subject has been carried out within the framework of the Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks‐Cloud Classification System (PERSIANN‐CCS). A new method, PIRSO (Probabilistic QPE using InfraRed Satellite Observations), is proposed to advance the use of uncertainty as an integral part of QPE. Probability distributions of precipitation rates are computed instead of deterministic values using a model quantifying the relation between satellite infrared brightness temperatures and the corresponding “true” precipitation rate. Ensembles of brightness temperatures‐to‐precipitation rate relationships are derived at a 30 min/0.04° scale. This approach conditions probabilistic quantitative precipitation estimates (PQPE) on the precipitation rate and typology. PIRSO's components were estimated based on a data sample covering two warm seasons over the conterminous USA. Precipitation probability maps outperform the deterministic PERSIANN‐CCS QPE. PIRSO is shown to mitigate systematic biases from deterministic retrievals, quantify uncertainty, and advance the monitoring of precipitation extremes. It also provides the basis for precipitation probability maps and satellite precipitation ensembles needed for satellite multi‐sensor merging of precipitation, early warning and mitigation of hydrometeorological hazards, and hydrological modelling.