Search

Remote sensing precipitation data have the characteristics of wide coverage and revealing spatiotemporal information, but their spatial resolution is low. The accuracy of the data is obviously different in different study areas and hydrometeorological conditions. This study evaluated four precipitation products in the Yellow River basin from 2001 to 2019, constructed the optimal combined product, conducted downscaling with various machine algorithms, and performed corrections using meteorological station precipitation data to analyze the spatiotemporal trends of precipitation. The results showed that (1) GPM and MSWEP had the best four evaluation indicators, with R2 values of 0.93 and 0.90, respectively, and the smallest FSE and RMSE, with a BIAS close to 0. A high-precision mixed precipitation dataset, GPM-MSWEP, was constructed. (2) Among the three methods, the downscaling results of DFNN showed higher accuracy. (3) The results, after correction with GWR, could more effectively enhance the accuracy of the data. (4) Precipitation in the Yellow River Basin showed a decreasing trend in January, September, and December, while it exhibited an increasing trend in other months and seasons, with 2002 and 2016 being points of abrupt change. This study provides a reference for the production of high-precision satellite precipitation products in the Yellow River basin. [ABSTRACT FROM AUTHOR]

Modeling inland water quality by remote sensing has already demonstrated its capacity to make accurate predictions. However, limitations still exist for applicability in diverse regions, as well as to retrieve non-optically active parameters (nOAC). Models are usually trained only with water samples from individual or local groups of waterbodies, which limits their capacity and accuracy in predicting parameters across diverse regions. This study aims to increase data availability to understand the performance of models trained with heterogeneous databases from both remote sensing and field measurement sources to improve machine learning training. This paper seeks to build a dataset with worldwide lake characteristics using data from water monitoring programs around the world paired with harmonized data of Landsat-8 and Sentinel-2. Additional feature engineering is also examined. The dataset is then used for model training and prediction of water quality at the global scale, time series analysis and water quality maps for lakes in different continents. Additionally, the modeling performance of nOACs are also investigated. The results show that trained models achieve moderately high correlations for SDD, TURB and BOD (R2 = 0.68) but lower performances for TSM and NO3-N (R2 = 0.43). The extreme learning machine (ELM) and the random forest regression (RFR) demonstrate better performance. The results indicate that ML algorithms can process remote sensing data and additional features to model water quality at the global scale and contribute to address the limitations of transferring and retrieving nOAC. However, significant limitations need to be considered, such as calibrated harmonization of water data and atmospheric correction procedures. Moreover, further understanding of the mechanisms that facilitate nOAC prediction is necessary. We highlight the need for international contributions to global water quality datasets capable of providing extensive water data for the improvement of global water monitoring. [ABSTRACT FROM AUTHOR]

Scarce water resources present a major hindrance to ensuring food security. Crop water productivity (WP), embraced as one of the Sustainable Development Goals (SDGs), is playing an integral role in the performance-based evaluation of agricultural systems and securing sustainable food production. This study aims at developing a cloud-based model within the Google Earth Engine (GEE) based on Landsat -7 and -8 satellite imagery to facilitate WP mapping at regional scales (30-m resolution) and analyzing the state of the water use efficiency and productivity of the agricultural sector as a means of benchmarking its WP and defining local gaps and targets at spatiotemporal scales. The model was tested in three major agricultural districts in the Lake Urmia Basin (LUB) with respect to five crop types, including irrigated wheat, rainfed wheat, apples, grapes, alfalfa, and sugar beets as the major grown crops. The actual evapotranspiration (ET) was estimated using geeSEBAL based on the Surface Energy Balance Algorithm for Land (SEBAL) methodology, while for crop yield estimations Monteith's Light Use Efficiency model (LUE) was employed. The results indicate that the WP in the LUB is below its optimum targets, revealing that there is a significant degree of work necessary to ameliorate the WP in the LUB. The WP varies between 0.49–0.55 (kg/m3) for irrigated wheat, 0.27–0.34 for rainfed wheat, 1.7–2.2 for apples, 1.2–1.7 for grapes, 5.5–6.2 for sugar beets, and 0.67–1.08 for alfalfa, which could be potentially increased up to 80%, 150%, 76%, 83%, 55%, and 48%, respectively. The spatial variation of the WP and crop yield makes it feasible to detect the areas with the best and poorest on-farm practices, thereby facilitating the better targeting of resources to bridge the WP gap through water management practices. This study provides important insights into the status and potential of WP with possible worldwide applications at both farm and government levels for policymakers, practitioners, and growers to adopt effective policy guidelines and improve on-farm practices. [ABSTRACT FROM AUTHOR]

Evaluating the accuracy of precipitation products is essential for many applications. The traditional method for evaluation is to calculate error metrics of products with gauge measurements that are considered as ground truth. The multiplicative triple collocation (MTC) method has been demonstrated powerful in error quantification of precipitation products when ground truth is not known. This study applied MTC to evaluate five precipitation products in Germany: two raw satellite-based products (CMORPH and PERSIANN), one reanalysis product (ERA-Interim), one soil moisture–based product (SM2RAIN-ASCAT), and one gauge-based product (REGNIE). Evaluation was performed at the 0.5° daily spatial–temporal scales. MTC involves a log transformation of data, necessitating dealing with zero values in daily precipitation. Effects of 12 different strategies for dealing with zero values on MTC results were investigated. Seven different triplet combinations were tested to evaluate the stability of MTC. Results showed that different strategies for replacing zero values had considerable effects on MTC-derived error metrics, particularly for root-mean-square error (RMSE). MTC with different triplet combinations generated different error metrics for individual products. The MTC-derived correlation coefficient (CC) was more reliable than RMSE. It is more appropriate to use MTC to compare the relative accuracy of different precipitation products. Based on CC with unknown truth, MTC with different triplet combinations produced the same ranking of products as the traditional method. A comparison of results from MTC and the classic TC with additive error model showed the potential limitation of MTC in arid areas or dry time periods with a large ratio of zero daily precipitation. [ABSTRACT FROM AUTHOR]

Projected increases in temporal precipitation variability, including intra‐annual and interannual variability, will likely have important impacts on terrestrial ecosystem productivity. The direction and magnitude of these impacts and how they vary across biomes, however, remain largely uncertain. Here, we review published literature that investigated the effect of different characteristics of temporal precipitation variability on vegetation productivity. We first reviewed commonly used methods, including manipulation experiments, process‐based modeling, and data‐driven analysis, and further discussed their strengths and limitations. Then, we summarized state‐of‐the‐art research on this topic by categorizing the results based on the characteristics of temporal precipitation variability. Given the same amount of growing season precipitation, a more extreme precipitation regime, characterized as fewer but larger precipitation events, tends to have a negative impact on vegetation productivity of most ecosystems except xeric grasslands and wet‐cold forests. Precipitation in the early growing season was found to be particularly important to vegetation productivity. Greater interannual precipitation variability tends to decrease vegetation productivity, but the reported patterns are complex, as both concave‐up and concave‐down precipitation‐productivity relations were found. Despite the progress made so far, critical challenges and knowledge gaps remain, such as the global‐scale impacts across different biomes, the role of biological adaption, and the contribution of individual precipitation events. Future research needs to combine manipulation experiments across a broad spectrum of ecosystem types and environmental gradients with model‐data integration strategies to disentangle the interactions between abiotic and biotic factors controlling vegetation responses to precipitation variability. This article is categorized under:Science of Water > Hydrological ProcessesWater and Life > Nature of Freshwater Ecosystems [ABSTRACT FROM AUTHOR]

Northeast China, the most important region for commercial grain farming in China, is vulnerable to drought due to high fluctuation in monthly rainfall. Timely, accurate and effective drought monitoring is very essential for securing the output of grain farming. In this study, three widely used drought indices were compared using satellite soil moisture data and their capability for drought monitoring was evaluated in Northeast China. Three indices are Normalized Monthly Precipitation Anomaly Percentage (NPA), Vegetation Health Index (VHI) and Normalized Vegetation Supply Water Index (NVSWI). In order to know the relationship between rainfall and drought especially over different local climate zones and give a more detail strategy for crop irrigation, a time-lag relationship is investigated based on the TRMM 3B43 monthly precipitation and NVSWI. Tendency rate (slope) was used to characterize the change of drought events in the region for the period 2001–2014. Results showed that: (1) the three selected indices were suitable for drought monitoring, with NVSWI performing the best in terms of the highest correlation with soil moisture, (2) an obvious time-lag was observed between rainfall and drought with the time lag being one month for all three climate zones, (3) drought occurred more frequently in spring and winter, while in summer drought occurred more easily in the west than in the east in Northeast China, (4) overall the frequency of drought was decreasing from 2001 to 2014 in Northeast China. [ABSTRACT FROM AUTHOR]

• First study to estimate Dissolved Organic Carbon in inland waters at a large scale. • The largest open source sample dataset AquaSat is used firstly for DOC estimation. • Satellite remote sensing and four different machine learning methods are used. • Including environmental data considerably improves the DOC estimation. • Gaussian Process Regression is the most promising method with its unique advantage. Dissolved Organic Carbon (DOC) in inland waters plays an essential role in the global carbon cycle and has significant public health effects. Machine learning (ML) together with remote sensing has emerged as a powerful and promising combination to quantify water quality parameters from space. However, inland water sample data for DOC is limited. Hence, little is known about the potential to quantify DOC content in inland waters, especially over large-scale areas. This study presents the first attempt to estimate DOC in inland waters over a large-scale area using satellite data and ML methods with the newly published open-source dataset AquaSat. Four ML approaches, namely Random Forest Regression (RFR), Support Vector Regression (SVR), Gaussian Process Regression (GPR), and a Multilayer Backpropagation Neural Network (MBPNN) were trained using more than 16 thousand samples across the continental United States matched with satellite data from Landsat 5, 7 and 8 missions. Satellite data from the Landsat missions were further extended with environmental data from the ERA5-Land product and used as input to train the ML algorithms. Our results show that including environmental data as inputs considerably improved the prediction of DOC for all ML algorithms, with GPR showing the most promising performance results with moderate estimation errors (RMSE: 4.08 mg/L). Permutation feature importance analysis showed that the wavelength range in the visible Green band (from Landsat) and the monthly average air temperature (from ERA5-Land) were the most important variables for the ML approaches. The results demonstrate the predictive strength of GPR and its useful feature to derive per pixel standard deviations for detailed analysis. Our results further highlight the important role of considering environmental processes to explain DOC variations over large scales. The application and performance of the GPR in mapping spatiotemporal variations of DOC in an entire water body were discussed by taking Lake Okeechobee (the 8th largest freshwater lake in the U.S.) as an illustrative example. While performance evaluation showed that DOC concentrations can be retrieved with adequate accuracy, algorithm development was challenged by the heterogenous nature of large-scale open source in situ data, issues related to atmospheric correction, and the low spatial and temporal resolution of the environmental predictors. This research demonstrates how open source, large-scale datasets like AquaSat in combination with ML and satellite remote sensing can make research toward large-scale estimation of inland water DOC more realistic while highlighting its remaining limitations and challenges. [Display omitted] [ABSTRACT FROM AUTHOR]

This study conducted a comprehensive evaluation of three satellite precipitation products (TRMM (Tropical Rainfall Measuring Mission) 3B42, CMORPH (the Climate Prediction Center (CPC) Morphing algorithm), and PERSIANN (Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks)) using data from 52 rain gauge stations over the Meichuan watershed, which is a representative watershed of the Poyang Lake Basin in China. All the three products were compared and evaluated during a 9-year period at different spatial (grid and watershed) and temporal (daily, monthly, and annual) scales. The results showed that at daily scale, CMORPH had the best performance with coefficients of determination (R2) of 0.61 at grid scale and 0.74 at watershed scale. For precipitation intensities larger than or equal to 25 mm, RMSE% of CMORPH and TRMM 3B42 were less than 50%, indicating CMORPH and TRMM 3B42 might be useful for hydrological applications at daily scale. At monthly and annual temporal scales, TRMM 3B42 had the best performances, with high R2 ranging from 0.93 to 0.99, and thus was deemed to be reliable and had good potential for hydrological applications at monthly and annual scales. PERSIANN had the worst performance among the three products at all cases. [ABSTRACT FROM AUTHOR]

The increasing competition for water resources requires a better understanding of flows, fluxes, stocks, and the services and benefits related to water consumption. This paper explains how public domain Earth Observation data based on Moderate Resolution Imaging Spectroradiometer (MODIS), Second Generation Meteosat (MSG), Tropical Rainfall Measurement Mission (TRMM) and various altimeter measurements can be used to estimate net water production (rainfall (P) > evapotranspiration (ET)) and net water consumption (ET > P) of Nile Basin agro-ecosystems. Rainfall data from TRMM and the Famine Early Warning System Network (FEWS-NET) RainFall Estimates (RFE) products were used in conjunction with actual evapotranspiration from the Operational Simplified Surface Energy Balance (SSEBop) and ETLook models. Water flows laterally between net water production and net water consumption areas as a result of runoff and withdrawals. This lateral flow between the 15 sub-basins of the Nile was estimated, and partitioned into stream flow and non-stream flow using the discharge data. A series of essential water metrics necessary for successful integrated water management are explained and computed. Net water withdrawal estimates (natural and humanly instigated) were assumed to be the difference between net rainfall (Pnet) and actual evapotranspiration (ET) and some first estimates of withdrawals-without flow meters-are provided. Groundwater-dependent ecosystems withdraw large volumes of groundwater, which exceed water withdrawals for the irrigation sector. There is a strong need for the development of more open-access Earth Observation databases, especially for information related to actual ET. The fluxes, flows and storage changes presented form the basis for a global framework to describe monthly and annual water accounts in ungauged river basins. [ABSTRACT FROM AUTHOR]

Remote Sensing, as a driver for water management decisions, needs further integration with monitoring water quality programs, especially in developing countries. Moreover, usage of remote sensing approaches has not been broadly applied in monitoring routines. Therefore, it is necessary to assess the efficacy of available sensors to complement the often limited field measurements from such programs and build models that support monitoring tasks. Here, we integrate field measurements (2013–2019) from the Mexican national water quality monitoring system (RNMCA) with data from Landsat-8 OLI, Sentinel-3 OLCI, and Sentinel-2 MSI to train an extreme learning machine (ELM), a support vector regression (SVR) and a linear regression (LR) for estimating Chlorophyll-a (Chl-a), Turbidity, Total Suspended Matter (TSM) and Secchi Disk Depth (SDD). Additionally, OLCI Level-2 Products for Chl-a and TSM are compared against the RNMCA data. We observed that OLCI Level-2 Products are poorly correlated with the RNMCA data and it is not feasible to rely only on them to support monitoring operations. However, OLCI atmospherically corrected data is useful to develop accurate models using an ELM, particularly for Turbidity (R2 = 0.7). We conclude that remote sensing is useful to support monitoring systems tasks, and its progressive integration will improve the quality of water quality monitoring programs. [ABSTRACT FROM AUTHOR]

The rapid expansion of oil palm is a major driver of deforestation and other associated damage to the climate and ecosystem in tropical regions, especially Southeast Asia. It is therefore necessary to precisely detect and monitor oil palm plantations to safeguard the ecosystem services and biodiversity of tropical forests. Compared with optical data, which are vulnerable to cloud cover, the Sentinel-1 dual-polarization C-band synthetic aperture radar (SAR) acquires global observations under all weather conditions and times of day and shows good performance for oil palm detection in the humid tropics. However, because accurately distinguishing mature and young oil palm trees by using optical and SAR data is difficult and considering the strong dependence on the input parameter values when detecting oil palm plantations by employing existing classification algorithms, we propose an innovative method to improve the accuracy of classifying the oil palm type (mature or young) and detecting the oil palm planting area in Sumatra by fusing Landsat-8 and Sentinel-1 images. We extract multitemporal spectral characteristics, SAR backscattering values, vegetation indices, and texture features to establish different feature combinations. Then, we use the random forest algorithm based on improved grid search optimization (IGSO-RF) and select optimal feature subsets to establish a classification model and detect oil palm plantations. Based on the IGSO-RF classifier and optimal features, our method improved the oil palm detection accuracy and obtained the best model performance (OA = 96.08% and kappa = 0.9462). Moreover, the contributions of different features to oil palm detection are different; nevertheless, the optimal feature subset performed the best and demonstrated good potential for the detection of oil palm plantations. [ABSTRACT FROM AUTHOR]

Remote-sensing-based machine learning approaches for water quality parameters estimation, Secchi Disk Depth (SDD) and Turbidity, were developed for the Valle de Bravo reservoir in central Mexico. This waterbody is a multipurpose reservoir, which provides drinking water to the metropolitan area of Mexico City. To reveal the water quality status of inland waters in the last decade, evaluation of MERIS imagery is a substantial approach. This study incorporated in-situ collected measurements across the reservoir and remote sensing reflectance data from the Medium Resolution Imaging Spectrometer (MERIS). Machine learning approaches with varying complexities were tested, and the optimal model for SDD and Turbidity was determined. Cross-validation demonstrated that the satellite-based estimates are consistent with the in-situ measurements for both SDD and Turbidity, with R2 values of 0.81 to 0.86 and RMSE of 0.15 m and 0.95 nephelometric turbidity units (NTU). The best model was applied to time series of MERIS images to analyze the spatial and temporal variations of the reservoir's water quality from 2002 to 2012. Derived analysis revealed yearly patterns caused by dry and rainy seasons and several disruptions were identified. The reservoir varied from trophic to intermittent hypertrophic status, while SDD ranged from 0–1.93 m and Turbidity up to 23.70 NTU. Results suggest the effects of drought events in the years 2006 and 2009 on water quality were correlated with water quality detriment. The water quality displayed slow recovery through 2011–2012. This study demonstrates the usefulness of satellite observations for supporting inland water quality monitoring and water management in this region. [ABSTRACT FROM AUTHOR]

A drone-based system for monitoring of laser-induced fluorescence from the aquatic environment was constructed. Fixed-range remote-sensing demonstration measurements were performed, and field recordings of natural river water fluorescence, oil-slicks as well as dye-marked natural water volumes were taken at drone flying heights of about 10 m. Our fluorosensor, weighing only 1.5 kg, and carried by a commercial drone, illustrates how airborne remote sensing based on fluorescence can be made cost-effective and readily applicable, while presently only in ambient low-light-level conditions. [ABSTRACT FROM AUTHOR]

As an important part of the farmland ecosystem, crop residues provide a barrier against water erosion, and improve soil quality. Timely and accurate estimation of crop residue coverage (CRC) on a regional scale is essential for understanding the condition of ecosystems and the interactions with the surrounding environment. Satellite remote sensing is an effective way of regional CRC estimation. Both optical remote sensing and microwave remote sensing are common means of CRC estimation. However, CRC estimation based on optical imagery has the shortcomings of signal saturation in high coverage areas and susceptibility to weather conditions, while CRC estimation using microwave imagery is easily influenced by soil moisture and crop types. Synergistic use of optical and microwave remote sensing information may have the potential to improve estimation accuracy. Therefore, the objectives of this study were to: (i) Analyze the correlation between field measured CRC and satellite derived variables based on Sentinel-1 and Sentinel-2, (ii) investigate the relationship of CRC with new indices (OCRI-RPs) which combine optical crop residues indices (OCRIs) and radar parameters (RPs), and (iii) to estimate CRC in Yucheng County based on OCRI-RPs by optimal subset regression. The correlations between field measured CRC and satellite derived variables were evaluated by coefficient of determination (R2) and root mean square error (RMSE). The results showed that the normalized difference tillage index (NDTI) and radar indices 2 (RI2) had relatively higher correlations with field measured CRC in OCRIs and RPs (R2 = 0.570, RMSE = 6.560% and R2 = 0.430, RMSE = 7.052%, respectively). Combining OCRIs with RPs by multiplying each OCRI with each RP could significantly improve the ability of indices to estimate CRC, as NDTI × RI2 had the highest R2 value of 0.738 and lowest RMSE value of 5.140%. The optimal model for CRC estimation by optimal subset regression was constructed by NDI71 × σ V V 0 and NDTI × σ V H 0 , with a R2 value of 0.770 and a RMSE value of 4.846%, which had a great improvement when compared with the best results in OCRIs and RPs. The results demonstrated that the combination of optical remote sensing information and microwave remote sensing information could improve the accuracy of CRC estimation. [ABSTRACT FROM AUTHOR]

Lake Tana is the largest lake in Ethiopia, and its lake outflow is the source of the Blue Nile River that provides vital water resources for many livelihoods and downstream/international stakeholders. Therefore, it is essential to quantify and monitor the water balance of Lake Tana. However, Lake Tana is poorly gauged, with more than 50% of Lake Tana Basin being ungauged from in-situ measurements, making it difficult to quantify the lake inflow from surrounding basins. The lack of in-situ measurements highlights the need for the innovative application of satellite remote sensing. This study explores how freely accessible satellite remote sensing can be used to complement routine weather data to quantify the water balance of Lake Tana and its surrounding catchments. This study particularly investigates whether the outflow from Lake Tana can be estimated with sufficient accuracy as the residual of the lake water balance. Monthly inflow into lake was computed as the total runoff from the surrounding catchments; the runoff was estimated as the residual of the land-based catchment water balance using satellite precipitation improved with an integrated downscaling-calibration procedure, satellite evapotranspiration, and a correction term for changes in land total storage (soil moisture storage and deep percolation). The outflow from Lake Tana was estimated as the residual of lake water balance by combining satellite-based lake precipitation, changes in water storage, and lake inflow with estimated lake evaporation. Evaluation using limited available measurements showed that estimated annual runoff for two gauged subbasins agreed well with measurements, with differences within 4%. The estimated annual outflow from Lake Tana was also close to measured outflow, with a difference of 12%. However, the estimated monthly runoff from catchments and monthly lake outflow were unsatisfactory, with large errors. [ABSTRACT FROM AUTHOR]

Lake ice is a sensitive indicator of climate change. Based on the disparities between the brightness temperatures of lake ice and water, passive microwave data can be used to monitor the ice variations of a lake. With focus on the analysis of long time series variability of lake ice, this study extracts four characteristic dates related to lake ice (the annual freeze start, freeze completion, ablation start and ablation completion dates) for Qinghai Lake from 1979 to 2016 using Scanning Multichannel Microwave Radiometer (SMMR) and Special Sensor Microwave Imager (SSM/I) passive microwave brightness temperature data. The corresponding freezing duration, ablation duration, complete freezing duration and ice coverage duration are calculated. Applying Moderate Resolution Imaging Spectroradiometer (MODIS) daily snow products, the accuracy of the results derived from passive microwave data is validated. The validation analysis shows a strong agreement (R 2 ranges from 0.70 to 0.85, mean absolute error (MAE) ranges from 2.25 to 3.94 days) in the freeze start, ablation start, and ablation completion dates derived from the MODIS data and passive microwave data; the ice coverage duration also has a small error (relative error (RE) = 2.95%, MAE = 3.13 days), suggesting that the results obtained from passive microwave data are reliable. The results show that the freezing dates of Qinghai Lake have been delayed and the ablation dates have advanced. Over 38 years, the freeze start date and freeze completion date have been pushed back by 6.16 days and 2.27 days, respectively, while the ablation start date and ablation completion date have advanced by 11.24 days and 14.09 days, respectively. The freezing duration and ablation duration have shortened by 3.89 days and 2.85 days, respectively, and the complete freezing duration and ice coverage duration have shortened by 14.84 days and 21.21 days, respectively. There is a significant negative correlation between the ice coverage duration and the mean air temperature in winter. [ABSTRACT FROM AUTHOR]

Multi-source merging is an established tool for improving large-scale precipitation estimates. Existing merging frameworks typically use gauge-based precipitation error statistics and neglect the inter-dependence of various precipitation products. However, gauge-observation uncertainties at daily and sub-daily time scales can bias merging weights and yield sub-optimal precipitation estimates, particularly over data-sparse regions. Likewise, frameworks ignoring inter-product error cross-correlation will overfit precipitation observation noise. Here, a Statistical Uncertainty analysis-based Precipitation mERging framework (SUPER) is proposed for addressing these challenges. Specifically, a quadruple collocation analysis is employed to estimate precipitation error variances and covariances for commonly used precipitation products. These error estimates are subsequently used for merging all products via a least-squares minimization approach. In addition, false-alarm precipitation events are removed via a reference rain/no-rain time series estimated by a newly developed categorical variable merging method. As such, SUPER does not require any rain gauge observations to reduce daily random and rain/no-rain classification errors. Additionally, by considering precipitation product inter-dependency, SUPER avoids overfitting measurement noise present in multi-source precipitation products. Results show that the overall RMSE of SUPER-based precipitation is 3.35 mm/day and the daily correlation with gauge observations is 0.71 [−] – metrics that are generally superior to recent precipitation reanalyses and remote sensing products. In this way, we seek to propose a new framework for robustly generating global precipitation datasets that can improve land surface and hydrological modeling skill in data-sparse regions. • We develop a new framework for global precipitation merging. • This framework leverages inter-product error dependency to avoid overfitting. • This framework minimizes the use of gauge-based precipitation observations. • This framework is theoretically more robust in data-sparse regions. [ABSTRACT FROM AUTHOR]

An exhaustive evaluation of water resources is a prerequisite for evidence-informed planning and implementing sustainable management strategies. However, the lack of sufficient information on water supply and consumption, alongside the technical limitations on comprehensive accounting for inter-relations and interactions between the subsystems, has resulted in decisions with often long-lasting outcomes and undesirable consequences. Water accounting is a tool for systematic acquisition, analysis, and reporting of water-related information to fill the existing knowledge gap on water flows and fluxes. In this study, Water Accounting Plus (WA+) framework is applied to the western region of Lake Urmia, a dying hyper-saline lake, to assess water use and crop water productivity (CWP) from 2010 to 2016. Remotely sensed information along with a distributed hydrological model (SPHY) is used to fill the information gap on water resources and inform effective policy actions. Our analysis reveals that the agricultural sector is neither productive nor efficient while there is a considerable scope to ameliorate water productivity and beneficial water use by adopting proper water management practices. Average CWP values for wheat, sugarbeet, vineyard, and apple vary between 0.38 and 0.55, 5.1–5.6, 1.5–1.7 and 1.9–2.3 (kg/m3), respectively while storage changes show consistent depletion, especially during dry year, up to 117 MCM. The results indicate that a 40% reduction in blue water use is achievable to supply additional water to revive Lake Urmia. This study highlights the importance of water accounting and information flow for decision-makers, practitioners, and farming communities to define practical targets and enhance productivity in water-scarce regions. [Display omitted] • WA+ was applied to the western region of Lake Urmia using RS and SPHY model. • Water accounts facilitate decision making and sustainable water management. • Water supply, availability, and productivity was analyzed. • There is significant potential to close CWP gap and reduce blue water use. • The proposed approach is valuable in data-scarce regions. [ABSTRACT FROM AUTHOR]

• Stepwisely including snow and glacier significantly improved runoff simulation. • Internal variable validation benefits model realism testing in data scarce basin. • One year calibration on runoff was sufficient to reproduce snow and glacier changes. • Single station precipitation as forcing outperformed Theissen average. Model realism is of vital importance in science of hydrology, in terms of realistic representation of hydrological processes and reliability of future prediction. Here, we employed a stepwise modeling approach that leverages flexible model structures and multi-source observations for robust streamflow simulation and internal variables validation with improved model realism. This framework is demonstrated in Yigong Zangbu River (YZR) basin, a data scarce glacier basin in the upper Brahmaputra River. We designed six experiments (Exp1–6) to use modeling as a tool to understand hydrological processes in this remote cold basin with extremely high altitude. In Exp1, we started with a distributed rainfall-runoff model (FLEXD) - representing the case that snow and glacier processes were ignored. Then, we stepwisely added snow and glacier processes into FLEXD, denoted as FLEXD-S (Exp2) and FLEXD-SG (Exp3), respectively, and such improvement of model structure led to significantly improved streamflow estimates. To explore the impact of different precipitation forcing on model performance, FLEXD-SG was driven by Theissen average (Exp3) and three individual stations' precipitation (Exp4–6). The model realism was tested by observed hydrograph, snow cover area (SCA) and glacier mass balance (GMB). Results showed that a robust and realistic hydrological modeling system was achieved in Exp6. In this modeling study, we learned that: 1) stepwise modeling is effective in investigating catchment behavior, and snow and glacier melting are the dominant hydrological processes in the YZR basin; 2) internal variables validation is beneficial to test model realism in data scarce basin; 3) the FLEXD-SG model calibrated by only one year hydrograph is sufficient to reproduce snow and glacier variations; 4) precipitation of a single station as forcing data could outperform Theissen average; 5) based on the well tested model configuration in Exp6, we analyzed simulated results, and reconstructed the long term hydrography (1961–2013), to support the potential competence for decision making on water resources management in practice. The proposed framework may significantly improve our skills in hydrological modeling over data-poor regions. [ABSTRACT FROM AUTHOR]

• An extended double instrumental variable (EIVD) algorithm is proposed. • EIVD can fully parameterize remote sensing hydrological product's error statistics. • EIVD is tested using numerical and real-world precipitation error analysis. • EIVD requires less independent datasets and is suitable for global applications. • EIVD can be used for optimally merging multi-source hydrological data. Optimally using multi-source remote-sensing (RS) and/or reanalyzed hydrological products requires knowledge of each product's accuracy and inter-product error cross-correlations. Quadruple collocation (QC) analysis can potentially solve for this error information without the reliance of high-quality ground references. However, QC requires at least three independent products for a variable of interest. At the global scale, obtaining three independent products is often a challenge. To address this issue, this study proposes an extended double instrumental variable algorithm (denoted as EIVD), which can accurately estimate product error and inter-product error cross-correlations using only two independent products – a requirement easier to meet in practice. Synthetic numerical experiments demonstrate that EIVD is robust and unbiased – provided product error auto-correlations are not strongly contrasting. The performance of EIVD is further tested via a (real-data) global precipitation error analysis using traditional QC results as a validation reference. The global consistency (i.e., spatial correlation) of QC- and EIVD-estimated product-truth correlation is above 0.86 [–] for all precipitation products being considered, and the relative mean difference of QC- and EIVD-based correlations is, on average, less than 5%. The spatial consistency of QC- and EIVD-based inter-product error cross-correlation is 0.47 [–] with a relative bias of 8%. A quantitative analysis demonstrates that regions with inconsistent EIVD and QC results are likely attributable to the violation of the QC error independency assumptions. Given the robustness of EIVD in fully parameterizing hydrological product error information, it is expected to improve the accuracy and efficiency of multi-source hydrological data merging and data assimilation. [ABSTRACT FROM AUTHOR]