Next Article in Journal
Carbon Pricing Strategies and Policies for a Unified Global Carbon Market
Previous Article in Journal
Ozone Pollution Impairs Athletic Performance in Female Football Players: A Gender-Specific Analysis
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Atmosphere
Volume 16
Issue 7
10.3390/atmos16070835
atmosphere-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Review

Review of Utilisation Methods of Multi-Source Precipitation Products for Flood Forecasting in Areas with Insufficient Rainfall Gauges

by
Yanhong Dou
1,2,
Ke Shi
1,
Hongwei Cai
3,
Min Xie
1,2 and
Ronghua Liu
1,2,*
1
State Key Laboratory of Water Cycle and Water Security, China Institute of Water Resources and Hydropower Research, Beijing 100038, China
2
Research Center on Flood & Drought Disaster Prevention and Reduction of the Ministry of Water Resources, Beijing 100038, China
3
School of Infrastructure Engineering, Dalian University of Technology, Dalian 116024, China
*
Author to whom correspondence should be addressed.
Atmosphere 2025, 16(7), 835; https://doi.org/10.3390/atmos16070835
Submission received: 25 May 2025 / Revised: 30 June 2025 / Accepted: 7 July 2025 / Published: 9 July 2025
(This article belongs to the Section Meteorology)
Browse Figures
Versions Notes

Abstract

The continuous release of global precipitation products offers a stable data source for flood forecasting in areas without rainfall gauges. However, due to constraints of forecast timeliness, only no/short-lag precipitation products can be utilised for flood forecasting, but these products are prone to significant errors. Therefore, the keys of flood forecasting in areas lacking rainfall gauges are selecting appropriate precipitation products, improving the accuracy of precipitation products, and reducing the errors of precipitation products by combination with hydrology models. This paper first presents the current no/short-lag precipitation products that are continuously updated online and for which the download of long series historical data is supported. Based on this, this paper reviews the utilisation methods of multi-source precipitation products for flood forecasting in areas with insufficient rainfall gauges from three perspectives: methods for precipitation product performance evaluation, multi-source precipitation fusion methods, and methods for coupling precipitation products with hydrological models. Finally, future research priorities are summarized: (i) to construct a quantitative evaluation system that can take into account both the accuracy and complementarity of precipitation products; (ii) to focus on the improvement of the areal precipitation fields interpolated by gauge-based precipitation in multi-source precipitation fusion; (iii) to couple real-time correction of flood forecasts and multi-source precipitation; and (iv) to enhance global sharing and utilization of rain gauge–radar data for improving the accuracy of satellite-based precipitation products.

1. Introduction

Precipitation is a critical driver in flood forecasting, directly affecting its accuracy [1]. The key methods for estimating precipitation include ground-based rain gauges, ground-based radar, satellite-based remote sensors, and numerical weather models [2]. Currently, rainfall gauges provide the most precise precipitation measurements. However, these measurements only represent the immediate vicinity of the gauges, necessitating a dense network to capture the true rainfall field. The World Meteorological Organization (WMO) [3] recommends a coverage of 250 km2 per gauge in mountainous regions and 900 km2 per gauge in plains. Nonetheless, in certain areas, factors such as terrain, geography, and economic constraints hinder the establishment and maintenance of rainfall gauge networks [4]. Consequently, these regions often suffer from a scarcity or sparsity of rainfall gauges, leading to inadequate measurement of rainfall data and limited spatial representativeness. This considerably impedes the development of flood control in the basin [5].
Compared to precipitation measurements from rain gauges and ground-based radar, those obtained via satellite remote sensing and numerical models offer global precipitation estimates and can be disseminated as precipitation products on major aerospace and meteorological websites [6,7]. Such products mainly include simulation products, analysis products and re-analysis products based on numerical models, as well as real-time products, near-real-time products, and post-processing products based on satellite remote sensing [8]. By 2024, hundreds of global precipitation products were available, with historical precipitation data spanning 10 to 20 years, meeting the informational needs for hydrological model calibration. These products provide a novel approach for flood forecasting in regions lacking rainfall gauges or with sparse gauge coverage [9].
The timeliness of precipitation measurements directly affects flood forecast speed. Precipitation products are categorised by acquisition time: real-time products with lag less than 1 h, near-real-time products with lag under 12 h, standard products, post-processing products, and re-analysis products with delays from 3 days to 6 months. Due to the timeliness requirements for flood forecasting, only highly prompt precipitation products are suitable for areas with insufficient rainfall gauges [1]. However, no-lag and short-lag precipitation products (e.g., real-time or near-real-time products, hereafter referred to as no/short-lag products) suffer from more significant errors than post-processing and reanalysis products due to the fact that they are not corrected by measured precipitation or the simplified steps of the precipitation retrieval algorithm [10,11,12]. In case of using no/short lag time precipitation products to force the hydrological model for flood forecasting, significant rainfall errors will increase the input uncertainty of the hydrological model, affect the water balance simulated by the hydrological model, and reduce the applicability of the forecast results [13]. Thus, a comprehensive understanding of precipitation product performance is crucial for flood forecasting in such regions. Selecting appropriate precipitation products, improving precipitation estimation, and enabling hydrological models to accommodate precipitation errors are crucial for accurate flood forecasting in areas with insufficient rainfall gauges.

2. Multisource No/Short-Lag Precipitation Products

Currently, no/short-lag precipitation products that are continuously updated online and for which the download of long series historical data is supported include GSMaP-V6-NRT (GSMaP-N) [14] and GSMaP-V6-Gauge-NRT (GSMaP-GN) [14] from the Japan Aerospace Exploration Agency (JAXA); IMERG-V6-Early (IMERG-E) [14] and IMERG-V6-Late (IMERG-L) [14] from the National Aeronautics and Space Administration (NASA); PERSIANN-CCS [15] from the University of California, Irvine (UCI); FY4A/FY4B QPE [16] from the National Satellite Meteorological Centre (NSMC); and analysis precipitation products from 10 countries or regions worldwide in the TIGGE dataset [17], such as the GRAPES from China Meteorological Administration, ECMWF from European Centre for Medium-Range Weather Forecasts, and JRSM from Japan Meteorological Agency. Precipitation products with no/short lag are categorised into three types based on their primary signal sources and retrieval algorithms: multi-sensor joint remote sensing precipitation products, visible/infrared remote sensing precipitation products, and meteorological model precipitation products. Multi-sensor joint remote sensing precipitation products [18] effectively compensate for the limitations of individual sensors by fusing various observational data, thereby enhancing the accuracy of precipitation estimation. However, the spatiotemporal resolution of such precipitation products remains limited. Visible/infrared remote sensing precipitation products [19] offer high spatial resolution and global coverage, making them suitable for large-scale continuous monitoring. Nevertheless, due to the relatively short wavelength and weak penetration capability of this sensor type, it can only detect atmospheric radiation above the cloud top, thereby exhibiting significant limitations in estimating rainfall beneath thick cloud layers. Meteorological model precipitation products [20], based on atmospheric physical processes, can provide detailed information on precipitation distribution and trend predictions. However, their results are highly sensitive to initial conditions and climate variability, leading to greater uncertainty during extreme weather events. The accuracy of precipitation products varies with different lag times. Pan et al. [21] analysed the overall accuracy of three precipitation products with different lag times. IMERG-E has a lag time of approximately four hours, IMERG-L about twelve hours, and IMERG-F around three months. Among them, IMERG-F demonstrates the best overall accuracy, while IMERG-E shows the lowest accuracy. The performance of precipitation products varies across different time scales. Zhang et al. [22] investigated the performance of gridded precipitation products and found that at the hourly scale, the overall annual performance of the SPEs surpasses that of the RPEs, with IMERG exhibiting the best performance. At the daily scale, MERRA has the best overall performance and CFSR shows great improvement over the hourly scale. These classifications are summarised in Table 1.

3. Methods for Precipitation Product Performance Evaluation

A comprehensive understanding of precipitation product performance is crucial for assessing their suitability. Precipitation estimation models exhibit substantial differences in their driving factors, modelling principles, and operational scenarios, resulting in varying performances. Extensive research into precipitation product performance presents systematised evaluation indicators, methods, and analytical perspectives. The accuracy evaluation indicators for precipitation products can be grouped into two categories: continuity indicators for quantitative precipitation estimation and classification indicators for qualitative estimation. Continuity indicators include the Pearson correlation coefficient (CC), root mean square error (RMSE), relative error (RE), mean error (ME), and the more comprehensive Kling–Gupta efficiency (KGE). Classification indicators include the probability of detection (POD), false alarm ratio (FAR), and the more comprehensive critical success index (CSI). In addition to the grid-based statistical evaluation methods mentioned above, object-based evaluation approaches have gained increasing attention in recent years. For instance, the SAL (Structure-Amplitude-Location) method [23] assesses precipitation by analysing the structure, intensity, and positional features of precipitation objects. This approach provides a more comprehensive evaluation of the spatial distribution characteristics of simulated precipitation, particularly in extreme rainfall event analysis, effectively compensating for the limitations of traditional grid-point statistical methods.
To comprehensively assess precipitation product characteristics, multiple indicators are typically employed simultaneously. For a more intuitive comparison of different precipitation products’ overall performance, two common methods are used. The first method integrates multiple evaluation indicators into a single metric using Euclidean distance [24], where a smaller distance from the optimal value signifies superior product quality. The second method utilises graphical tools such as Taylor diagrams [25] and performance diagrams [26]. Taylor diagrams plot indicators like RMSE and CC, illustrating the quantitative performance of precipitation products on a two-dimensional plane. In contrast, performance diagrams display indicators such as critical success index and probability of detection, representing the qualitative classification performance on a two-dimensional plane. Tang et al. [27] combined Taylor diagrams and performance diagrams to evaluate and compare both quantitative and qualitative performance of eleven different precipitation products across various regions and seasons.
These evaluation indicators require computations at locations where measured values serve as the baseline. Therefore, it is difficult to analyse the quality and performance of precipitation products in areas without rainfall gauges, as well as the performance of precipitation products in monitoring areal precipitation and spatial distribution of precipitation in areas with sparse rainfall gauges. Two primary methods are utilised to address this issue. The first is the triple collocation (TC) method [28] and its enhanced variants [29,30]. These approaches estimate the RMSE and CC of each product by analysing pairwise covariances among three mutually independent precipitation datasets. Roebeling et al. [31] were the first to apply the TC method to assess the performance of precipitation products. The second method is hydrological evaluation [32]. This technique compares the simulated discharge of hydrological models driven by precipitation products against observed discharge, indirectly assessing the spatiotemporal accuracy of precipitation products using metrics such as the flood peak error and Nash–Sutcliffe efficiency of discharge series. Additionally, it assesses the hydrological utility of precipitation products in regions with abundant rainfall data. In summary, the performance evaluation of precipitation products should adopt different methodologies based on data availability conditions (as shown in Figure 1). For areas with ground-based rainfall measurements, assessment can be conducted using continuous metrics, categorical metrics, or comprehensive composite indicators. For regions lacking rain gauge data, evaluation should rely on either the triple collocation method or hydrological evaluation approaches.

4. Methods for Multisource Precipitation Fusion

Although precipitation data measured by gauge are the most accurate, their precision and coverage decreases beyond their range of measurement. Precipitation products offer global-scale estimations, but only reflect the general precipitation trend. Moreover, the strengths and weaknesses of multisource precipitation products vary under different conditions. Consequently, integrating multisource precipitation has emerged as a research hotspot for improving precipitation estimation accuracy.
(1)
Fusion of gauge-measured precipitation and single-source precipitation products
A common fusion approach for gauge-measured precipitation and a single-source precipitation product is to correct the precipitation product against the gauge-measured precipitation (hereafter referred to as the “correction approach”). In the 1990s, following the launch of the Tropical Rainfall Measuring Mission (TRMM) satellite, the Global Precipitation Climatology Project (GPCP) initially sought to integrate satellite-based precipitation products with gauge-based measurements [33], resulting in a global monthly precipitation dataset with a spatial resolution of 0.25°. Subsequently, research on using gauge-based precipitation measurements to correct discrepancies in precipitation products has grown substantially. The method for correcting precipitation products involves several steps: first, at grid points with rainfall gauges, error patterns in the precipitation products are identified using gauge measurements as references; then, for grid points without rainfall gauges, these error patterns are used to estimate precipitation product errors, allowing for the adjustment of the precipitation field derived from the product. Therefore, precipitation product correction methods can be categorised into two types based on their focus.
The first category prioritises quantifying precipitation product error patterns by utilising long-term precipitation data to calibrate or train a correction model at grid points with rainfall gauges, which is subsequently applied to grid points without rainfall gauges. Representative methods include (non)linear regression analysis [34,35], quantile mapping [36,37], geographic variation analysis [38,39,40] and machine learning [41,42]. As illustrated in Route a of Figure 2, the error field obtained through training with these representative methods is fused with the precipitation product’s estimated rainfall field to generate the corrected precipitation estimates. For example, Lu et al. [43] employed precipitation data from over a thousand rainfall gauges in the Tianshan region to establish a conversion relationship between measured precipitation and IMERG products using stepwise regression and geographic weighting methods, integrating various terrain factors and vegetation indicators as correlated variables. This conversion was then applied to grid points in areas lacking rainfall gauges. The results revealed that both correction methods significantly enhanced IMERG precipitation estimates, with the geographic weighting method outperforming the traditional stepwise regression approach. Similarly, Tang et al. [44] refined IMERG precipitation estimates in the upper basin using long and short-term memory (LSTM) networks, utilising gauge-measured rainfall in the downstream basin. This correction markedly improved the accuracy of flood forecasting in small mountainous basins.
The second category addresses the spatial expansion of precipitation product errors, which involves estimating the error field of the precipitation product, and subsequently adjusting the precipitation product accordingly. Prominent methods in this category comprise the objective analysis method [45], optimal interpolation methods [46,47], dual kernel smoothing [48], and kriging interpolation method [35]. As illustrated in Route b of Figure 2, the error field estimated by these representative methods is fused with the precipitation product’s rainfall field to produce the corrected precipitation estimates. Yan et al. [49] assessed the spatial error field of IMERG by evaluating product discrepancies at grid points using rainfall gauge data and kriging interpolation within the Hanjiang River Basin. By overlaying the error field onto IMERG, they significantly enhanced the accuracy of daily precipitation measurements in the basin. Baez-Villanueva et al. [50] summarised the key steps of these methodologies and their application impacts across different regions through comparative tables.
(2)
Fusion of gauge-measured precipitation and multisource precipitation products
There is considerable potential for beneficial integration among precipitation products from various sources [51,52]. In the 1990s, Xie and Arkin [53] pioneered the integration of gauge-measured precipitation with multisource precipitation products. They combined data from gauges, satellites, and re-analysis products to create the widely utilised global monthly precipitation dataset CMAP. As precipitation products proliferated, the fusion of gauge-measured data with multisource precipitation products (hereafter referred to as the “fusion approach”) has become the leading method for improving precipitation estimation [54,55]. Similar to the correction approach, the fusion approach operates on the model principle of “training at grid with gauges—applying at grid without gauges.” Compared to the correction approach, the fusion approach involves more data sources and also needs to consider the relationship between each data source, including the influencing factors such as terrain, vegetation, and meteorology. Therefore, additional correlated variables must be accounted for during fusion, leading to increased algorithmic complexity or rendering the algorithm a black box.
Fusion methods are primarily categorised into two types. The first type includes weighted averaging techniques, such as inverse RMSE weighting [56], optimal weighting [54], and Bayesian model averaging [57]. As illustrated in Route a of Figure 3, these representative methods are employed to train and determine the fusion weights at locations with rain gauges, followed by the integration of multi-source precipitation products at ungauged sites. Zhu et al. [56] employed a morphologically adaptive approach to identify rainfall events in both temporal and spatial dimensions. They integrated multisource precipitation data at grid points lacking rainfall gauges by using the reciprocal of the root mean square error of precipitation products against gauges as weights. Similarly, Hong et al. [58] determined fusion weights of each grid using an artificial neural network. The covariates in the neural network included elevation, barometric pressure, and wind speed at each grid point, resulting in fusion weights that vary spatially and temporally.
The second category consists of machine learning-based fusion methods such as random forest [50,59], artificial neural networks [60], long short-term memory [61], and convolutional neural networks [62,63]. As depicted in Route b of Figure 3, these representative methods are used to train models at gauged locations, after which the trained models are applied to estimate precipitation at ungauged sites. For instance, Baez-Villanueva et al. [50] trained a model using precipitation estimates from multiple sources at grid points as correlated variables, using gauge-measured precipitation as the benchmark. They subsequently applied this trained model to estimate precipitation at grid points lacking rainfall gauges. Zhang et al. [52] implemented a fusion strategy termed “classification first and then regression”, where the random forest classifier initially distinguished between no-rain and rainy periods. This was followed by regression functions from four machine learning algorithms. Wu et al. [61] integrated convolutional neural networks (CNNs) with LSTM networks. Initially, CNNs extracted spatial features of precipitation from various sources; thereafter, LSTM networks captured the temporal dependencies of these spatial features. Following fusion, precipitation accuracy improved by 15–17% compared to the original product.
Additionally, some studies have proposed a two-stage approach that involves deviation correction of precipitation products from each source, followed by the fusion of the corrected multisource precipitation. For instance, Ma et al. [64] first corrected each precipitation product by a generalised regression function between product error and topography before fusing the corrected precipitation products using Bayesian model averaging. This two-stage methodology successfully mitigated interference from product outliers, and the fused precipitation accuracy improved by approximately 10% compared to the original products.
(3)
Fusion of multisource precipitation products in areas without rainfall gauges
Regions without rainfall gauges lack measured benchmarks necessary for training or calibrating models that correct or fuse multisource precipitation products. Consequently, implementing conventional fusion methods is problematic. Common approaches include model transplantation, fusion techniques based on triple collocation, and leveraging other hydrological variables to adjust precipitation products.
Model transplantation entails training a fusion model for multisource precipitation data using gauge-measured precipitation at regions with rainfall gauges, then applying it to regions without gauges. For example, Wang et al. [65] developed a CNN-based algorithm for jointly inverting precipitation using multisource infrared signals. To address the limited availability of gauge-measured precipitation, the model was initially pre-trained in the United States, where rainfall gauges are plentiful, and subsequently transferred to mainland China. The precipitation estimates from this approach significantly outperformed the similar PERSIANN-CCS.
The triple collocation method assesses the errors among a trio of precipitation products by analysing pairwise covariances among three mutually independent products, then fuses the multisource precipitation data by accounting for the error of each product using techniques such as linear regression, non-linear regression, or weighted averaging [66]. Chen et al. [67] introduced a multisource precipitation product fusion approach utilising the multiplicative triple collocation (MTC) method. Validation in the Yangtze River Basin demonstrated that this approach matched the accuracy of traditional fusion methods in case of not taking gauge-measured precipitation into account. To address the constraint of the conventional triple collocation method, which is limited to combining three precipitation product types, Xu et al. [68] developed an expanded triple-collocation-based fusion technique that integrated 13 monthly and 11 daily precipitation products globally.
Precipitation is a crucial component of the water cycle, which is closely linked to hydrological variables such as soil moisture content and runoff. Consequently, observations of precipitation-sensitive hydrological elements can guide the correction or integration of precipitation products, including soil moisture data. By employing satellite-based soil moisture content, Massari et al. [69] improved the accuracy of precipitation products and established the initial wet conditions of basins for flood simulations. Utilising data assimilation with satellite-derived soil moisture estimates, Roman-Cascon et al. [70] refined three precipitation products: CMORPH, TMPA, and PERSIANN. However, this type of approach is not applicable in areas with complex terrain, dense forests and other strong radio frequency interference due to the limitations of the remote sensing equipment for soil moisture content. Discharge data are easily accessible in flood-prone regions. Si et al. [71] introduced the dynamic system response curve method, which corrected rainfall estimates using observed discharge mediated by hydrological models. Nonetheless, this approach depends on a lumped hydrological model, limiting it to correcting only the average areal precipitation of the basin.
Studies exploring the above fusion methods (Figure 4) meticulously assess algorithm performance across different months, locations, terrains, elevations, precipitation intensities, and climatic conditions, showcasing their broad applicability. However, few studies have focused on the effect of rainfall gauge density on the performance of fusion methods [39,47,52]. Bai et al. [39] analysed the performance of corrected PERSIANN-CRD by using gauge-measured rainfall at varying gauge densities. They discovered that when a relatively high number of rainfall gauges (3000~5000 km2/gauge) is used to correct the precipitation products, the accuracy is inferior to that of rainfall interpolated by measured rainfall only, even though the corrected rainfall is a substantial improvement over the original precipitation products; at this point, the correction method is no longer of practical value.

5. Methods for Coupling Precipitation Products with Hydrological Models

Precipitation products are crucial for flood forecasting in areas lacking precipitation data, and errors in precipitation products are difficult to completely correct. Therefore, mitigating model input uncertainty originating from errors in precipitation products by integrating them with hydrological models is a key focus in flood forecasting using multisource precipitation data [72,73]. Precipitation estimates with considerable uncertainty used as inputs to hydrological models lead to improved flood forecasting accuracy through enhancements to the model input, the hydrological model itself, and intermediate variables. Therefore, this section reviews flood forecasting research utilising multisource precipitation products from these three perspectives.
For model inputs, particularly enhanced precipitation estimation, Section 4 explores methods for improving precipitation estimates by fusing multisource data, which are largely independent of the flood forecasting process. Three prevalent coupling methods improve the integration of enhanced precipitation estimates with hydrological models. The first method entails multisource precipitation data fusion models. Specifically, training or calibrating a multisource precipitation data fusion model can optimise precipitation merging for more accurate precipitation-driven flood forecasting. The second approach addresses hydrological model parameters. Specifically, using either original precipitation products or enhanced estimates to drive hydrological models and calibrating model parameters to improve flood forecasting accuracy can adjust the parameters to account for systematic errors in precipitation estimation. For example, Yuan et al. [74] recalibrated hydrological model parameters using IMERG and TRMM as inputs. After recalibration, precipitation products showed greater hydrological application potential compared to flood forecasting based on original parameters. However, their accuracy in flood forecasting when driven by gauge-measured precipitation remained limited. The third approach involves simultaneously training a multisource precipitation data fusion model and calibrating hydrological model parameters. For instance, Chaing et al. [75] treated the fusion weights of multisource precipitation products as parameters to be optimised. Aiming to minimise simulated discharge errors, they optimised product fusion weights alongside hydrological model parameters. In the first two methods, adjusting hydrological model parameters to accommodate uncertainties in precipitation products caused the parameters to deviate from their physically meaningful ranges, resulting in poor scalability of the hydrological models [76,77].
For hydrological models, in addition to recalibrating hydrological model parameters, certain studies have adopted a forecasting strategy that replaced physical models with data-driven models such as machine learning [78,79]. Data-driven models can directly simulate discharges using multisource precipitation estimates, which carry significant uncertainties, whilst bypassing intermediate processes such as precipitation product correction, multisource data fusion, runoff generation, and discharge concentration through opaque black-box simulations. These models partially alleviate the operational challenges associated with coupled forecasting [78]. For example, Kumar et al. [80] utilised measured precipitation from sparse gauges, near-real-time precipitation estimates from TRMM, and soil moisture content estimates from ASCAT as input variables for discharge forecasting. They replaced the traditional hydrological physical model with a support vector machine approach to reduce operational complexity. Across various training schemes, the accuracy of flood forecasting experienced moderate improvements. Similarly, Nanda et al. [81] compared the discharge forecasting capabilities of artificial neural networks, wavelet neural networks, and other machine learning models in regions lacking precipitation data. They exclusively employed short sequences of near-real-time TRMM precipitation estimates as input variables for discharge forecasting and substituted initial soil moisture content with cumulative precipitation from the preceding period, as estimated by precipitation products. The results demonstrated robust flood forecasting capabilities. Jiang et al. [82] first extracted the temporal and spatial features of precipitation products and leaf area indices using a machine vision model, and then trained the artificial neural network with these features as the correlated variables for discharge forecasting. The research demonstrated satisfactory performance in short-term prediction, long-term simulation, and transfer learning. Notable differences were observed in the training methods of machine learning models, the spatiotemporal application of multisource precipitation datasets, and the selection of correlated variables in this research design.
Soil moisture content is the primary calibration target for intermediate variables in hydrological simulations because precipitation data, which often include a notable degree of uncertainty, directly influence soil moisture simulations used to drive hydrological models. Soil moisture promptly responds to precipitation and directly affects flood event volumes; hence, it plays a crucial role as a key state variable with an integrative role [83]. Calibrating the soil moisture content generated by hydrological models typically involves real-time soil moisture measurements or discharge data. However, soil moisture gauges are scarce in regions with limited precipitation data [84]. Consequently, due to the lack of soil moisture observations, calibration studies frequently utilise remote sensing-based soil moisture products instead of in situ gauge measurements [69]. For example, Chen et al. [85] calibrated a hydrological model driven by near-real-time TRMM precipitation data using soil moisture estimates from ASCAT and SMOS through ensemble Kalman filtering. The application of this technique in 13 watersheds in the central United States showed that this method can simulate discharge well. Currently, two methods utilise remote sensing soil moisture products to aid flood forecasting with multisource precipitation data. The first method calibrates multisource precipitation products using remote sensing soil moisture data before driving hydrological models, as detailed in Section 4 [70]. The second method involves real-time calibration of the soil state within hydrological models, using remote sensing soil moisture content alongside multisource precipitation inputs [85]. Massari et al. [86] compared the effectiveness of these approaches, finding that the first is more suitable for predicting high discharge, whereas the second excels in forecasting low discharge. Hydrological gauges are typically established in target flood forecasting areas, providing more readily available reliable discharge measurements than soil moisture data. Lee et al. [87] applied variational data assimilation, using measured discharge to adjust the soil moisture content simulated by the SAC-SMA hydrological model driven by the CMORPH precipitation product. Hydrological models driven by precipitation products calibrated with measured discharge demonstrate higher forecasting accuracy than flood forecasting corrections based on soil moisture content observations, due to the superior observation accuracy of measured discharge [88].

6. Summary and Outlook

The continuous release of global precipitation products offers a stable data source for flood forecasting in areas without rainfall gauges. However, due to timeliness constraints, only no/short-lag precipitation products can be utilised for flood forecasting, and these products are prone to significant errors. Therefore, selecting appropriate precipitation products, enhancing precipitation estimation, and integrating precipitation products with hydrological models are essential for flood forecasting in regions lacking rainfall gauges. This study initially presents the current no/short-lag precipitation products that are continuously updated online and support the download of extensive historical data. Subsequently, it reviews research advancements in flood forecasting for regions without rainfall gauges from three perspectives: methods for evaluating precipitation product performance, multisource precipitation data fusion techniques, and the integration of precipitation products with hydrological models. Finally, this study draws the following conclusions:
(1)
Quantitative evaluation and analysis of precipitation product performance is the cornerstone of precipitation product selection. Although numerous accuracy evaluation metrics exist, relying solely on accuracy is limited by substantial errors in no-lag or short-lag precipitation products. As precipitation products diversify, selecting a product with marginally lower accuracy but unique error characteristics can significantly enhance product diversity. This selection promotes the complementary strengths and weaknesses necessary for multisource product fusion. Current single-source accuracy indicators qualitatively assess the strengths and weaknesses of each precipitation product by separately evaluating their accuracy under various conditions. Therefore, as multisource product fusion increasingly supplants the independent use of single-source products, the interaction performance among multisource precipitation products warrants attention alongside individual product accuracy.
(2)
Integrating measured precipitation from gauges with multisource precipitation products is essential for improved precipitation estimation. Existing multisource data fusion methods primarily utilise gauge measurements to enhance precipitation products, with precipitation interpolation at sparse gauges only benefiting cases with extremely limited gauge coverage. Current fusion methods lack practical applicability. Compared to the recommendations of the World Meteorological Organization of 250–900 km2 per gauge, the gauge network at the specified threshold density remains exceedingly sparse, potentially constraining global-scale precipitation observations. Accurately and consistently monitoring the spatial distribution of precipitation at the basin scale is challenging, thereby hindering effective flood forecasting and necessitating further enhancements through the use of precipitation products. Enhancing the efficacy of precipitation estimation via gauge interpolation is a promising area for future research.
(3)
However, it is difficult to fully correct errors in precipitation products. Integrating multisource precipitation data with hydrological models and managing input uncertainties are crucial for flood forecasting in regions with limited precipitation data. Real-time correction of hydrological models, driven by precipitation products and informed by measured discharge, can dynamically refine forecasts using observed data, thereby incorporating high-precision information into flood forecasting where data are scarce. Therefore, this integration approach is emerging as a key research direction to enhance the capacity of hydrological models to mitigate precipitation product errors. Nonetheless, the errors in no/short-lag precipitation products should not be overlooked. Comprehensive integration of multisource precipitation data with real-time flood forecasting adjustments to improve prediction accuracy while mitigating correction volatility is expected to be a key research focus in the future.
(4)
The integration of rain gauge and radar data plays a critical role in enhancing the accuracy of precipitation products. Current precipitation estimates primarily rely on satellite remote sensing retrievals, which exhibit systematic deviations compared to ground-based observations. High-quality rain gauge–radar merged datasets from regions such as South Korea (KMA) and Japan (AMeDAS) combine ground station measurements with three-dimensional radar observations, providing high-precision reference data with a spatial resolution of 1 km and a temporal resolution of 10 min. These datasets are instrumental in correcting biases in satellite-derived precipitation estimates. However, their accessibility remains restricted to specific regions, limiting their global applicability. Moving forward, it is imperative to promote international data sharing, establish unified standards, and develop advanced data fusion techniques to improve precipitation monitoring and flood forecasting capabilities in data-sparse regions.

Author Contributions

Conceptualization, Y.D. and K.S.; methodology, Y.D. and K.S.; formal analysis, Y.D. and H.C.; investigation, Y.D., K.S., and M.X.; writing—original draft preparation, Y.D.; writing—review and editing, K.S., H.C., M.X., and R.L.; visualization, Y.D. and H.C.; supervision, R.L.; funding acquisition, R.L. All authors have read and agreed to the published version of the manuscript.

Funding

This study was funded by the National Key Research and Development Program of China (2023YFC3006700) and Ningbo Municipal Water Resources Science and Technology Program Project (NSKA202507).

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Li, J.; Li, L.; Zhang, T.; Xing, H.; Shi, Y.; Li, Z.; Wang, C.; Liu, J. Flood forecasting based on radar precipitation nowcasting using U-net and its improved models. J. Hydrol. 2024, 632, 130871. [Google Scholar] [CrossRef]
  2. Michaelides, S.; Levizzani, V.; Anagnostou, E.; Bauer, P.; Kasparis, T.; Lane, J.E. Precipitation: Measurement, remote sensing, climatology and modeling. Atmos. Res. 2009, 94, 512–533. [Google Scholar] [CrossRef]
  3. World Meteorological Organization. Guide to Hydrological Practices: Hydrology-From Measurement to Hydrological Information; World Meteorological Organization: Geneva, Switzerland, 2008. [Google Scholar]
  4. Bayat, B.; Hosseini, K.; Nasseri, M.; Karami, H. Challenge of rainfall network design considering spatial versus spatiotemporal variations. J. Hydrol. 2019, 574, 990–1002. [Google Scholar] [CrossRef]
  5. Toté, C.; Patricio, D.; Boogaard, H.; van der Wijngaart, R.; Tarnavsky, E.; Funk, C. Evaluation of Satellite Rainfall Estimates for Drought and Flood Monitoring in Mozambique. Remote Sens. 2015, 7, 1758–1776. [Google Scholar] [CrossRef]
  6. Kang, X.; Dong, J.; Crow, W.T.; Wei, L.; Zhang, H. The Conditional Bias of Extreme Precipitation in Multi-Source Merged Data Sets. Geophys. Res. Lett. 2024, 51, e2024GL111378. [Google Scholar] [CrossRef]
  7. José, G.-C.; Viera, R.; Sergio, S.-G.; Félix, F. Evaluation of ERA5 and ERA5-Land reanalysis precipitation datasets over Spain (1951–2020). Atmos. Res. 2023, 284, 106606. [Google Scholar]
  8. Ray, R.L.; Sishodia, R.P.; Tefera, G.W. Evaluation of gridded precipitation data for hydrologic modeling in north-central Texas. Remote Sens. 2022, 14, 3860. [Google Scholar] [CrossRef]
  9. Li, J.; Yuan, D.; Ma, M.; Liu, J. Simulation of karst floods with a hydrological model improved by meteorological model coupling. J. Hydrometeorol. 2022, 23, 185–207. [Google Scholar] [CrossRef]
  10. Zhao, H.; Yang, B.; Yang, S.; Huang, Y.; Dong, G.; Bai, J.; Wang, Z. Systematical estimation of GPM-based global satellite mapping of precipitation products over China. Atmos. Res. 2018, 201, 206–217. [Google Scholar] [CrossRef]
  11. Navarro, A.; Merino, A.; Sánchez, J.L.; Kummerow, C.; Tapiador, F.J. Assessment of IMERG Precipitation Estimates over Europe. Remote Sens. 2019, 11, 2470. [Google Scholar] [CrossRef]
  12. Mahmoud, M.T.; Hamouda, M.A.; Mohamed, M.M. Spatiotemporal evaluation of the GPM satellite precipitation products over the United Arab Emirates. Atmos. Res. 2019, 219, 200–212. [Google Scholar] [CrossRef]
  13. McMillan, H.; Jackson, B.; Clark, M.; Kavetski, D.; Woods, R. Rainfall uncertainty in hydrological modelling: An evaluation of multiplicative error models. J. Hydrol. 2011, 400, 83–94. [Google Scholar] [CrossRef]
  14. Hsu, J.; Huang, W.-R.; Liu, P.-Y. Performance assessment of GPM-based near-real-time satellite products in depicting diurnal precipitation variation over Taiwan. J. Hydrol. Reg. Stud. 2021, 38, 100957. [Google Scholar] [CrossRef]
  15. Hsu, J.; Huang, W.-R.; Liu, P.-Y. Comprehensive Analysis of PERSIANN Products in Studying the Precipitation Variations over Luzon. Remote Sens. 2022, 14, 5900. [Google Scholar] [CrossRef]
  16. Siyu, Z.; Ziqiang, M.; Songkun, Y.; Jintao, X.; Xiaoqing, L. PECAM-FY4A&B: Precipitation Estimation Using Chromatographic Analysis Method by Merging Enhanced Multispectral Infrared Observations From FengYun-4A and 4B. IEEE Trans. Geosci. Remote Sens. 2025, 63, 1–18. [Google Scholar]
  17. Medina, H.; Tian, D.; Srivastava, P.; Pelosi, A.; Chirico, G.B. Medium-range reference evapotranspiration forecasts for the contiguous United States based on multi-model numerical weather predictions. J. Hydrol. 2018, 562, 502–517. [Google Scholar]
  18. Guo, R.; Fan, X.; Zhou, H.; Liu, Y. Multi-Sensor Precipitation Estimation from Space: Data Sources, Methods and Validation. Remote Sens. 2024, 16, 4753. [Google Scholar] [CrossRef]
  19. Liu, Q.; Fu, Y. Comparison of radiative signals between precipitating and non-precipitating clouds in frontal and typhoon domains over East Asia. Atmos. Res. 2010, 96, 436–446. [Google Scholar] [CrossRef]
  20. Amini, S.; Azizian, A.; Arasteh, D.P. How reliable are TIGGE daily deterministic precipitation forecasts over different climate and topographic conditions of Iran? Meteorol. Appl. 2021, 28, e2013. [Google Scholar] [CrossRef]
  21. Pan, X.; Wu, H.; Chen, S.; Nergui, N.; Huang, Z.; Chen, W.; Li, C.; Li, X. Evaluation and Applicability Analysis of GPM Satellite Precipitation over Mainland China. Remote Sens. 2023, 15, 2866. [Google Scholar] [CrossRef]
  22. Zhang, L.; Chen, X.; Lai, R.; Zhu, Z. Performance of Satellite-Based and Reanalysis Precipitation Products under Multi-Temporal Scales and Extreme Weather in Mainland China. J. Hydrol. 2022, 605, 127389. [Google Scholar] [CrossRef]
  23. Hiraga, Y.; Tahara, R. Sensitivity of localized heavy rainfall in Northern Japan to WRF physics parameterization schemes. Atmos. Res. 2025, 314, 107802. [Google Scholar] [CrossRef]
  24. Chang, X.W.; Paige, C.C. Euclidean distances and least squares problems for a given set of vectors. Appl. Numer. Math. 2007, 57, 1240–1244. [Google Scholar] [CrossRef]
  25. Taylor, K.E. Summarizing multiple aspects of model performance in a single diagram. J. Geophys. Res.-Atmos. 2001, 106, 7183–7192. [Google Scholar] [CrossRef]
  26. Roebber, P.J. Visualizing Multiple Measures of Forecast Quality. Weather. Forecast. 2009, 24, 601–608. [Google Scholar] [CrossRef]
  27. Tang, G.; Clark, M.P.; Papalexiou, S.M.; Ma, Z.; Hong, Y. Have satellite precipitation products improved over last two decades? A comprehensive comparison of GPM IMERG with nine satellite and reanalysis datasets. Remote Sens. Environ. 2020, 240, 111697. [Google Scholar] [CrossRef]
  28. Stoffelen, A. Toward the true near-surface wind speed: Error modeling and calibration using triple collocation. J. Geophys. Res.-Ocean. 1998, 103, 7755–7766. [Google Scholar] [CrossRef]
  29. Alemohammad, S.H.; McColl, K.A.; Konings, A.G.; Entekhabi, D.; Stoffelen, A. Characterization of precipitation product errors across the United States using multiplicative triple collocation. Hydrol. Earth Syst. Sci. 2015, 19, 3489–3503. [Google Scholar] [CrossRef]
  30. McColl, K.A.; Vogelzang, J.; Konings, A.G.; Entekhabi, D.; Piles, M.; Stoffelen, A. Extended triple collocation: Estimating errors and correlation coefficients with respect to an unknown target. Geophys. Res. Lett. 2014, 41, 6229–6236. [Google Scholar] [CrossRef]
  31. Roebeling, R.A.; Wolters, E.L.A.; Meirink, J.F.; Leijnse, H. Triple Collocation of Summer Precipitation Retrievals from SEVIRI over Europe with Gridded Rain Gauge and Weather Radar Data. J. Hydrometeorol. 2012, 13, 1552–1566. [Google Scholar] [CrossRef]
  32. Zhang, L.; Xin, Z.H.; Zhou, H.C. Assessment of TMPA 3B42V7 and PERSIANN-CDR in Driving Hydrological Modeling in a Semi-Humid Watershed in Northeastern China. Remote Sens. 2020, 12, 3133. [Google Scholar] [CrossRef]
  33. Huffman, G.J.; Adler, R.F.; Arkin, P.; Chang, A.; Ferraro, R.; Gruber, A.; Janowiak, J.; McNab, A.; Rudolf, B.; Schneider, U. The Global Precipitation Climatology Project (GPCP) Combined Precipitation Dataset. Bull. Am. Meteorol. Soc. 1997, 78, 5–20. [Google Scholar] [CrossRef]
  34. Chen, H.; Bin, Y.; Gourley, J.J.; Wen, D.; Qi, W.; Yang, K. A Novel Real-Time Error Adjustment Method With Considering Four Factors for Correcting Hourly Multi-Satellite Precipitation Estimates. IEEE Trans. Geosci. Remote Sens. 2022, 60, 1–11. [Google Scholar] [CrossRef]
  35. Vernimmen, R.R.; Hooijer, A.; Aldrian, E.; Van Dijk, A.I. Evaluation and bias correction of satellite rainfall data for drought monitoring in Indonesia. Hydrol. Earth Syst. Sci. 2012, 16, 133–146. [Google Scholar] [CrossRef]
  36. Johnson, F.; Sharma, A. Accounting for interannual variability: A comparison of options for water resources climate change impact assessments. Water Resour. Res. 2011, 47, W04508. [Google Scholar] [CrossRef]
  37. Themessl, M.J.; Gobiet, A.; Heinrich, G. Empirical-statistical downscaling and error correction of regional climate models and its impact on the climate change signal. Clim. Change 2012, 112, 449–468. [Google Scholar] [CrossRef]
  38. Chen, Y.; Huang, J.; Sheng, S.; Mansaray, L.R.; Liu, Z.; Wu, H.; Wang, X. A new downscaling-integration framework for high-resolution monthly precipitation estimates: Combining rain gauge observations, satellite-derived precipitation data and geographical ancillary data. Remote Sens. Environ. 2018, 214, 154–172. [Google Scholar] [CrossRef]
  39. Bai, X.Y.; Wu, X.Q.; Wang, P. Blending long-term satellite-based precipitation data with gauge observations for drought monitoring: Considering effects of different gauge densities. J. Hydrol. 2019, 577, 124007. [Google Scholar] [CrossRef]
  40. Chao, L.; Zhang, K.; Li, Z.; Zhu, Y.; Wang, J.; Yu, Z. Geographically weighted regression based methods for merging satellite and gauge precipitation. J. Hydrol. 2018, 558, 275–289. [Google Scholar] [CrossRef]
  41. Gumindoga, W.; Rientjes, T.H.M.; Haile, A.T.; Makurira, H.; Reggiani, P. Performance of bias-correction schemes for CMORPH rainfall estimates in the Zambezi River basin. Hydrol. Earth Syst. Sci. 2019, 23, 2915–2938. [Google Scholar] [CrossRef]
  42. Yuhang, Z.; Aizhong, Y.; Phu, N.; Bita, A.; Soroosh, S.; Kuolin, H. QRF4P-NRT: Probabilistic Post-Processing of Near-Real-Time Satellite Precipitation Estimates Using Quantile Regression Forests. Water Resour. Res. 2022, 58, e2022WR032117. [Google Scholar]
  43. Xinyu, L.; Guoqiang, T.; Xiuqin, W.; Yan, L.; Lihong, J.; Guohui, X.; Shuai, L.; Yingxin, Z. Correcting GPM IMERG precipitation data over the Tianshan Mountains in China. J. Hydrol. 2019, 575, 1239–1252. [Google Scholar]
  44. Tang, X.; Yin, Z.; Qin, G.; Guo, L.; Li, H. Integration of Satellite Precipitation Data and Deep Learning for Improving Flash Flood Simulation in a Poor-Gauged Mountainous Catchment. Remote Sens. 2021, 13, 5083. [Google Scholar] [CrossRef]
  45. Rozante, J.R.; Moreira, D.S.; de Goncalves, L.G.G.; Vila, D.A. Combining TRMM and Surface Observations of Precipitation: Technique and Validation over South America. Weather Forecast. 2010, 25, 885–894. [Google Scholar] [CrossRef]
  46. Xie, P.P.; Xiong, A.Y. A conceptual model for constructing high-resolution gauge-satellite merged precipitation analyses. J. Geophys. Res.-Atmos. 2011, 116, D21106. [Google Scholar] [CrossRef]
  47. Wu, Z.; Zhang, Y.; Sun, Z.; Lin, Q.; He, H. Improvement of a combination of TMPA (or IMERG) and ground-based precipitation and application to a typical region of the East China Plain. Sci. Total Environ. 2018, 640, 1165–1175. [Google Scholar] [CrossRef]
  48. Li, M.; Shao, Q.X. An improved statistical approach to merge satellite rainfall estimates and raingauge data. J. Hydrol. 2010, 385, 51–64. [Google Scholar] [CrossRef]
  49. Yan, X.; Chen, H.; Tian, B.; Sheng, S.; Wang, J.; Kim, J.-S. A Downscaling-Merging Scheme for Improving Daily Spatial Precipitation Estimates Based on Random Forest and Cokriging. Remote Sens. 2021, 13, 2040. [Google Scholar] [CrossRef]
  50. Baez-Villanueva, O.M.; Zambrano-Bigiarini, M.; Beck, H.E.; McNamara, I.; Ribbe, L.; Nauditt, A.; Birkel, C.; Verbist, K.; Giraldo-Osorio, J.D.; Thinh, N.X. RF-MEP: A novel Random Forest method for merging gridded precipitation products and ground-based measurements. Remote Sens. Environ. 2020, 239, 111606. [Google Scholar] [CrossRef]
  51. Chen, S.; Xiong, L.; Ma, Q.; Kim, J.-S.; Chen, J.; Xu, C.-Y. Improving daily spatial precipitation estimates by merging gauge observation with multiple satellite-based precipitation products based on the geographically weighted ridge regression method. J. Hydrol. 2020, 589, 125156. [Google Scholar] [CrossRef]
  52. Zhang, L.; Li, X.; Zheng, D.; Zhang, K.; Ma, Q.; Zhao, Y.; Ge, Y. Merging multiple satellite-based precipitation products and gauge observations using a novel double machine learning approach. J. Hydrol. 2021, 594, 125969. [Google Scholar] [CrossRef]
  53. Xie, P.P.; Arkin, P.A. Global precipitation: A 17-year monthly analysis based on gauge observations, satellite estimates, and numerical model outputs. Bull. Am. Meteorol. Soc. 1997, 78, 2539–2558. [Google Scholar] [CrossRef]
  54. Beck, H.E.; Wood, E.F.; Pan, M.; Fisher, C.K.; Miralles, D.G.; van Dijk, A.I.J.M.; McVicar, T.R.; Adler, R.F. MSWEP V2 Global 3-Hourly 0.1 degrees Precipitation: Methodology and Quantitative Assessment. Bull. Am. Meteorol. Soc. 2019, 100, 473–502. [Google Scholar] [CrossRef]
  55. Beck, H.E.; van Dijk, A.I.J.M.; Levizzani, V.; Schellekens, J.; Miralles, D.G.; Martens, B.; de Roo, A. MSWEP: 3-hourly 0.25 degrees global gridded precipitation (1979-2015) by merging gauge, satellite, and reanalysis data. Hydrol. Earth Syst. Sci. 2017, 21, 589–615. [Google Scholar] [CrossRef]
  56. Zhu, S.; Ma, Z.; Xu, J.; He, K.; Liu, H.; Ji, Q.; Tang, G.; Hu, H.; Gao, H. A Morphology-Based Adaptively Spatio-Temporal Merging Algorithm for Optimally Combining Multisource Gridded Precipitation Products with Various Resolutions. IEEE Trans. Geosci. Remote Sens. 2022, 60, 1–21. [Google Scholar] [CrossRef]
  57. Ma, Y.; Yang, H.; Chen, Y.; Yuan, Y.; Tang, G.; Yao, Y.; Di, L.; Li, C.; Han, Z.; Liu, R. Performance of Optimally Merged Multisatellite Precipitation Products Using the Dynamic Bayesian Model Averaging Scheme Over the Tibetan Plateau. J. Geophys. Res.-Atmos. 2018, 123, 814–834. [Google Scholar] [CrossRef]
  58. Hong, Z.; Han, Z.; Li, X.; Di, L.; Tang, G.; Wang, J. Generation of an Improved Precipitation Dataset from Multisource Information over the Tibetan Plateau. J. Hydrometeorol. 2021, 22, 1275–1295. [Google Scholar]
  59. Nguyen, G.V.; Le, X.H.; Nguyen, V.L.; Jung, S.; Yeon, M.; Lee, G. Application of Random Forest Algorithm for Merging Multiple Satellite Precipitation Products across South Korea. Remote Sens. 2021, 13, 4033. [Google Scholar] [CrossRef]
  60. Turlapaty, A.C.; Anantharaj, V.G.; Younan, N.H.; Turk, F.J. Precipitation data fusion using vector space transformation and artificial neural networks. Pattern Recognit. Lett. 2010, 31, 1184–1200. [Google Scholar] [CrossRef]
  61. Wu, H.; Yang, Q.; Liu, J.; Wang, G. A spatiotemporal deep fusion model for merging satellite and gauge precipitation in China. J. Hydrol. 2020, 584, 124664. [Google Scholar] [CrossRef]
  62. Sadeghi, M.; Nguyen, P.; Hsu, K.; Sorooshian, S. Improving near real-time precipitation estimation using a U-Net convolutional neural network and geographical information. Environ. Model. Softw. 2020, 134, 104856. [Google Scholar] [CrossRef]
  63. Wang, C.G.; Tang, G.Q.; Gentine, P. PrecipGAN: Merging Microwave and Infrared Data for Satellite Precipitation Estimation Using Generative Adversarial Network. Geophys. Res. Lett. 2021, 48, e2020GL092032. [Google Scholar] [CrossRef]
  64. Ma, Y.; Sun, X.; Chen, H.; Yang, H.; Zhang, Y. A two-stage blending approach for merging multiple satellite precipitation estimates and rain gauge observations: An experiment in the northeastern Tibetan Plateau. Hydrol. Earth Syst. Sci. 2021, 25, 359–374. [Google Scholar] [CrossRef]
  65. Wang, C.; Tang, G.; Xiong, W.; Ma, Z.; Zhu, S. Infrared precipitation estimation using convolutional neural network for FengYun satellites. J. Hydrol. 2021, 603, 127113. [Google Scholar] [CrossRef]
  66. Lyu, F.; Tang, G.; Behrangi, A.; Wang, T.; Tan, X.; Ma, Z.; Xiong, W. Precipitation Merging Based on the Triple Collocation Method Across Mainland China. IEEE Trans. Geosci. Remote Sens. 2021, 59, 3161–3176. [Google Scholar] [CrossRef]
  67. Chen, C.; He, M.; Chen, Q.; Zhang, J.; Li, Z.; Wang, Z.; Duan, Z. Triple collocation-based error estimation and data fusion of global gridded precipitation products over the Yangtze River basin. J. Hydrol. 2022, 605, 127307. [Google Scholar] [CrossRef]
  68. Xu, L.; Chen, N.; Moradkhani, H.; Zhang, X.; Hu, C. Improving Global Monthly and Daily Precipitation Estimation by Fusing Gauge Observations, Remote Sensing, and Reanalysis Data Sets. Water Resour. Res. 2020, 56, e2019WR026444. [Google Scholar] [CrossRef]
  69. Massari, C.; Brocca, L.; Moramarco, T.; Tramblay, Y.; Didon-Lescot, J.-F. Potential of soil moisture observations in flood modelling: Estimating initial conditions and correcting rainfall. Adv. Water Resour. 2014, 74, 44–53. [Google Scholar] [CrossRef]
  70. Román-Cascón, C.; Pellarin, T.; Gibon, F.; Brocca, L.; Cosme, E.; Crow, W.; Fernández-Prieto, D.; Kerr, Y.H.; Massari, C. Correcting satellite-based precipitation products through SMOS soil moisture data assimilation in two land-surface models of different complexity: API and SURFEX. Remote Sens. Environ. 2017, 200, 295–310. [Google Scholar] [CrossRef]
  71. Si, W.; Gupta, H.V.; Bao, W.; Jiang, P.; Wang, W. Improved Dynamic System Response Curve Method for Real-Time Flood Forecast Updating. Water Resour. Res. 2019, 55, 7493–7519. [Google Scholar] [CrossRef]
  72. Sun, R.C.; Yuan, H.L.; Yang, Y.Z. Using multiple satellite-gauge merged precipitation products ensemble for hydrologic uncertainty analysis over the Huaihe River basin. J. Hydrol. 2018, 566, 406–420. [Google Scholar] [CrossRef]
  73. Falck, A.S.; Maggioni, V.; Tomasella, J.; Vila, D.A.; Diniz, F.L.R. Propagation of satellite precipitation uncertainties through a distributed hydrologic model: A case study in the Tocantins-Araguaia basin in Brazil. J. Hydrol. 2015, 527, 943–957. [Google Scholar] [CrossRef]
  74. Yuan, F.; Wang, B.; Shi, C.; Cui, W.; Zhao, C.; Liu, Y.; Ren, L.; Zhang, L.; Zhu, Y.; Chen, T.; et al. Evaluation of hydrological utility of IMERG Final run V05 and TMPA 3B42V7 satellite precipitation products in the Yellow River source region, China. J. Hydrol. 2018, 567, 696–711. [Google Scholar] [CrossRef]
  75. Chiang, Y.-M.; Hsu, K.-L.; Chang, F.-J.; Yang, H.; Sorooshian, S. Merging multiple precipitation sources for flash flood forecasting. J. Hydrol. 2007, 340, 183–196. [Google Scholar] [CrossRef]
  76. Jiang, S.; Ren, L.; Hong, Y.; Yong, B.; Yang, X.; Yuan, F.; Ma, M. Comprehensive evaluation of multi-satellite precipitation products with a dense rain gauge network and optimally merging their simulated hydrological flows using the Bayesian model averaging method. J. Hydrol. 2012, 452, 213–225. [Google Scholar] [CrossRef]
  77. Bitew, M.M.; Gebremichael, M. Evaluation of satellite rainfall products through hydrologic simulation in a fully distributed hydrologic model. Water Resour. Res. 2011, 47, W06526. [Google Scholar] [CrossRef]
  78. Bonakdari, H.; Zaji, A.H.; Binns, A.D.; Gharabaghi, B. Integrated Markov chains and uncertainty analysis techniques to more accurately forecast floods using satellite signals. J. Hydrol. 2019, 572, 75–95. [Google Scholar] [CrossRef]
  79. Akhtar, M.K.; Corzo, G.A.; van Andel, S.J.; Jonoski, A. River flow forecasting with artificial neural networks using satellite observed precipitation pre-processed with flow length and travel time information: Case study of the Ganges river basin. Hydrol. Earth Syst. Sci. 2009, 13, 1607–1618. [Google Scholar] [CrossRef]
  80. Kumar, A.; Ramsankaran, R.A.A.J.; Brocca, L.; Muñoz-Arriola, F. A simple machine learning approach to model real-time streamflow using satellite inputs: Demonstration in a data scarce catchment. J. Hydrol. 2021, 595, 126046. [Google Scholar] [CrossRef]
  81. Nanda, T.; Sahoo, B.; Beria, H.; Chatterjee, C. A wavelet-based non-linear autoregressive with exogenous inputs (WNARX) dynamic neural network model for real-time flood forecasting using satellite-based rainfall products. J. Hydrol. 2016, 539, 57–73. [Google Scholar] [CrossRef]
  82. Jiang, S.; Zheng, Y.; Babovic, V.; Tian, Y.; Han, F. A computer vision-based approach to fusing spatiotemporal data for hydrological modeling. J. Hydrol. 2018, 567, 25–40. [Google Scholar] [CrossRef]
  83. Meng, S.S.; Xie, X.H.; Liang, S.L. Assimilation of soil moisture and streamflow observations to improve flood forecasting with considering runoff routing lags. J. Hydrol. 2017, 550, 568–579. [Google Scholar] [CrossRef]
  84. Brocca, L.; Melone, F.; Moramarco, T.; Singh, V.P. Assimilation of Observed Soil Moisture Data in Storm Rainfall-Runoff Modeling. J. Hydrol. Eng. 2009, 14, 153–165. [Google Scholar] [CrossRef]
  85. Chen, F.; Crow, W.T.; Ryu, D. Dual Forcing and State Correction via Soil Moisture Assimilation for Improved Rainfall-Runoff Modeling. J. Hydrometeorol. 2014, 15, 1832–1848. [Google Scholar] [CrossRef]
  86. Massari, C.; Camici, S.; Ciabatta, L.; Brocca, L. Exploiting Satellite-Based Surface Soil Moisture for Flood Forecasting in the Mediterranean Area: State Update Versus Rainfall Correction. Remote Sens. 2018, 10, 292. [Google Scholar] [CrossRef]
  87. Lee, H.; Zhang, Y.; Seo, D.-J.; Xie, P. Utilizing satellite precipitation estimates for streamflow forecas ting via adjustment of mean field bias in precipitation data and assimilation of streamflow observations. J. Hydrol. 2015, 529, 779–794. [Google Scholar] [CrossRef]
  88. Si, W.; Bao, W.M.; Gupta, H.V. Updating real-time flood forecasts via the dynamic system response curve method. Water Resour. Res. 2015, 51, 5128–5144. [Google Scholar] [CrossRef]
Atmosphere 16 00835 g001
Figure 1. Classification structure of precipitation product performance evaluation methods.
Figure 1. Classification structure of precipitation product performance evaluation methods.
Atmosphere 16 00835 g001
Atmosphere 16 00835 g002
Figure 2. Algorithm flow chart summarizing the fusion of gauge-measured precipitation and single-source precipitation products, where ∆ P G i represents the product error at the grid point where the rain gauge is located; ∆ P x , y represents the product error at the grid point with coordinates (x, y); f (*) represents the method for calculating the error field based on different models; A → represents attribute of a grid point; ε represents the residual error; ∆ P → represents the error field; P S → represents the rainfall field of the product; P M → represents the fused rainfall field; a represents Route a; b represents Route b.
Figure 2. Algorithm flow chart summarizing the fusion of gauge-measured precipitation and single-source precipitation products, where ∆ P G i represents the product error at the grid point where the rain gauge is located; ∆ P x , y represents the product error at the grid point with coordinates (x, y); f (*) represents the method for calculating the error field based on different models; A → represents attribute of a grid point; ε represents the residual error; ∆ P → represents the error field; P S → represents the rainfall field of the product; P M → represents the fused rainfall field; a represents Route a; b represents Route b.
Atmosphere 16 00835 g002
Atmosphere 16 00835 g003
Figure 3. Algorithm flow chart summarizing the fusion of gauge-measured precipitation and multisource precipitation products, where P G i represents the fused rainfall at the grid point where the rain gauge is located; P x , y M represents the fused rainfall at the grid point with coordinates (x, y); P G i 1 represents the rainfall at the grid point of rain gauge i for product 1; P x , y 1 represents the rainfall at the grid point with coordinates (x, y) for product 1; f (*) represents the methods for different fusion products; ω represents the fusion weight; A → represents the attribute of a grid point; ε represents the residual error; a represents Route a; b represents Route b.
Figure 3. Algorithm flow chart summarizing the fusion of gauge-measured precipitation and multisource precipitation products, where P G i represents the fused rainfall at the grid point where the rain gauge is located; P x , y M represents the fused rainfall at the grid point with coordinates (x, y); P G i 1 represents the rainfall at the grid point of rain gauge i for product 1; P x , y 1 represents the rainfall at the grid point with coordinates (x, y) for product 1; f (*) represents the methods for different fusion products; ω represents the fusion weight; A → represents the attribute of a grid point; ε represents the residual error; a represents Route a; b represents Route b.
Atmosphere 16 00835 g003
Atmosphere 16 00835 g004
Figure 4. Classification structure of multisource precipitation fusion methods.
Figure 4. Classification structure of multisource precipitation fusion methods.
Atmosphere 16 00835 g004
Table 1. Classification of no/short-lag precipitation products.
Table 1. Classification of no/short-lag precipitation products.
CategoryPrimary Signal SourcesRetrieval AlgorithmsPrecipitation ProductsTime ResolutionAccessible Time
Multi-sensor joint remote sensingMicrowave, visible and infrared signals from satellite-based sensorsScattering and emission characteristics of microwaves by water vapour, and the relationship between cloud-top brightness temperature and precipitation rateGSMaP-N
GSMaP-GN
IMERG-E
IMERG-L		1 h/1 d
1 h/1 d
0.5 h/1 d
0.5 h/1 d		4-h lag
4-h lag
4-h lag
12-h lag
Visible/infrared
remote sensing		Visible and infrared signals from satellite-based sensorsRelationship between cloud-top brightness temperature and precipitation ratePERSIANN-CCS
FY4A/FY4B QPE		1 h/1 d
1 h/1 d		0.5-h lag
0.5-h lag
Meteorological model simulationMeteorological elements and fluxes from terrestrial, oceanic and upper-air observationsAtmospheric dynamics equationGRAPES
ECMWF
JRSM
……		6 h/1 d
6 h/1 d
6 h/1 d
……		/
/
/
……
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Dou, Y.; Shi, K.; Cai, H.; Xie, M.; Liu, R. Review of Utilisation Methods of Multi-Source Precipitation Products for Flood Forecasting in Areas with Insufficient Rainfall Gauges. Atmosphere 2025, 16, 835. https://doi.org/10.3390/atmos16070835

AMA Style

Dou Y, Shi K, Cai H, Xie M, Liu R. Review of Utilisation Methods of Multi-Source Precipitation Products for Flood Forecasting in Areas with Insufficient Rainfall Gauges. Atmosphere. 2025; 16(7):835. https://doi.org/10.3390/atmos16070835

Chicago/Turabian Style

Dou, Yanhong, Ke Shi, Hongwei Cai, Min Xie, and Ronghua Liu. 2025. "Review of Utilisation Methods of Multi-Source Precipitation Products for Flood Forecasting in Areas with Insufficient Rainfall Gauges" Atmosphere 16, no. 7: 835. https://doi.org/10.3390/atmos16070835

APA Style

Dou, Y., Shi, K., Cai, H., Xie, M., & Liu, R. (2025). Review of Utilisation Methods of Multi-Source Precipitation Products for Flood Forecasting in Areas with Insufficient Rainfall Gauges. Atmosphere, 16(7), 835. https://doi.org/10.3390/atmos16070835

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop