Search

• A spatial random forest-based downscaling and merging method (SRF-DM) is proposed. • SRF-DM performs better than the original satellite-based precipitation products. • SRF-DM is more accurate than the classical machine learning-based methods. • Spatial autocorrelation is the most important variable. • SRF-DM has a good performance even when the gauge density is low. Publicly available satellite precipitation products (SPPs) are essential inputs for hydrological models. However, the existing SPPs suffer from coarse resolutions and large uncertainties, which greatly hinder their widespread applications. Thus, to fully utilize the unique characteristics of each individual SPP, a spatial random forest (SRF)-based downscaling and merging (SRF-DM) method is proposed to fuse multiple SPPs and gauge observations in this paper. The proposed method incorporates spatial autocorrelation information between precipitation observations through a covariate. This covariate is estimated using ordinary kriging interpolation on spatial neighbors that have a high correlation with the target location. The proposed method was used to generate 1-km daily precipitation data using the datasets from five satellite precipitation products (IMERG, GSMaP, CHIRPS, MORPH, and PERSIANN) and gauge observations in the period from January 1, 2015 to December 31, 2019 over Sichuan province, China. The performance of SRF-DM was compared with the original SPPs, XGBoost-based downscaling and merging (XGB-DM), two double-layer machine learning (ML) methods, namely RF-RF and RF-artificial neural network (RF-ANN), and three single ML methods, namely RF, ANN, and RF-based Merging Procedure (RF-MEP). The experimental results demonstrate the following findings: (i) SRF-DM is more accurate than the original SPPs across different temporal scales (i.e., daily, monthly, and seasonal) in terms of statistical and categorical accuracy measures, (ii) SRF-DM outperforms the six ML-based methods in estimating high precipitation and capturing spatial precipitation patterns, (iii) the spatial autocorrelation between neighboring precipitation data is the most significant factor in both downscaling and merging, and (iv) the performance of SRF-DM is influenced by gauge density, but it still yields good results even with a limited number of gauges. On the whole, SRF-DM is considered an effective tool for generating daily precipitation products with high accuracy and resolution, which is primarily attributed to its satisfactory performance in regions with diverse precipitation patterns and topography. [ABSTRACT FROM AUTHOR]

• The proposed model improves the forecast skill for sub-daily precipitation. • The model was improved by treating precipitation forecasts as censored data. • The proposed model achieves similar forecast skill with an EMOS model. Statistical post-processing methods have been applied in hydrometeorological forecasting to correct the bias and spread error in raw forecasts. Among various post-processing methods, the meta-Gaussian distribution model (MGD) is one of the early successful methods for post-processing of precipitation forecasts and has been applied in the National Weather Service's Hydrologic Ensemble Forecast System (HEFS), together with the mix-type meta-Gaussian distribution model (MMGD). However, recent studies have shown that the original MGD cannot yield reliable forecasts especially for sub-daily precipitation forecasts (e.g., 6-hourly). In this paper, we improved the MGD model by applying the censored maximum likelihood estimation (CMLE) method. We conducted experiments using GEFS reforecasts in Huai river basin in China to evaluate its performance. The results show that the proposed method performs better than the original MGD for sub-daily precipitation forecasts. The proposed method also achieves similar forecast skill with the state-of-the-art censored, shifted Gamma distribution-based ensemble MOS (CSGD-EMOS) if both use ensemble mean as the only predictor. The results indicate that the proposed CMLE-MGD can be useful for further applications such as flood forecasting that needs forecasts of high temporal resolution. [ABSTRACT FROM AUTHOR]

Summary: In this paper, the trends of the annual streamflow and precipitation and cross correlations between them were analyzed in nine large river basins of China during 1956–2005. The results indicate that: (1) the annual mean streamflow decreases in arid and semi-arid regions of north China; however, increasing trends occur in south and Southwest China; (2) the annual streamflow and precipitation exhibit reasonable correlation in nine large river basins except those located in inland areas. The annual streamflow over most areas of China is fed by precipitation; however, the decline in streamflow is faster than the decreases of precipitation since 1970s in the arid and semi-arid regions of north China. The relationship between the annual precipitation and streamflow presents a non-stationary state in north China. This non-stationary relationship is strongly influenced by both human activities and precipitation changes; (3) a significant increase of water use might be the major factor responsible for the steeper decline in streamflow than in precipitation in Haihe River, Yellow River and Songliao River basins in north China. In inland river areas, increase of water use and actual evapotranspiration might result in decline in streamflow although precipitation has an increase tendency. This paper sheds light on the non-stationary relationship between annual precipitation and streamflow and possible underlying causes, which will be helpful for a better understanding of the changes of precipitation and streamflow in China at large scale and in other regions of the world. [Copyright &y& Elsevier]

Summary The climate elasticity of runoff is an important indicator that is used to quantify the relationship between changes in runoff and changes in climate variables. It is a function of both climate and catchment characteristics. Recently, Yang and Yang (2011) proposed an analytical derivation of climate elasticity (YY2011), in which a parameter n was used to represent the impact of the catchment characteristics. In China, both climate and catchment characteristics have large spatial variations. To understand the spatial variation of hydrologic response to climate change, this paper divided China into 210 catchments, further calculated the parameter n , and then estimated the climate elasticity and evaluated the contribution of climate change to runoff for each catchment. The results show that n ranges from 0.4 to 3.8 (with a mean of 1.3 and a standard deviation of 0.6), which has a logarithmic relationship with catchment slope; the precipitation elasticity ranges from 1.1 to 4.8 (with a mean of 1.9 and a standard deviation of 0.6), which shows a large regional variation, smaller values (1.1–2.0) mainly appearing in Southern China, the Songhua River basin and the Northwest, and larger values (2.1–4.8) mainly appearing in the Hai River basin, the Liao River basin and the Yellow River basin. In addition, climate contribution to runoff exhibits a large regional variation, the largest positive values (1.1–3.1%/a) occurring in the Northwest, the largest negative values (−1.0 to −0.5%/a) occurring in the Hai River basin and the middle reach of the Yellow River basin. In theory, the YY2011 method is a first-order approximation. The approximation underestimates the precipitation ( P ) contribution to runoff when P increases and overestimates that when P decreases, and the relative error has a median of ∼3% and a maximum of ∼20% when 10% precipitations change in those catchments of China. [ABSTRACT FROM AUTHOR]

Highlights: [•] This paper examines the impacts of rain gauge densities on runoff simulations. [•] Variation of rain gauge densities impacts the areal mean rainfall considerably. [•] The locations of rain gauges impact the model performances considerably. [•] Great impacts on model performances occurring when rain gauges fall below 0.4 station per 1000km2. [Copyright &y& Elsevier]

Summary: Oxygen and hydrogen stable isotopes (δ 18O and δD) in precipitation collected in every event from 2007 to 2009 in Guangzhou, South China, are presented in this paper. The total correlation between δ 18O and δD is obtained as δD =(8.46±0.13) δ 18O+(15.0±0.9). More negative δ 18O and δD generally occur during summer and autumn, while less negative or even positive δ 18O and δD occur during winter and spring. Significant negative correlations between precipitation δ 18O and temperature, and between precipitation δ 18O and precipitation amount are observed. Regression line changes from year to year are likely due to changes in moisture sources for the precipitation. The moisture contributed by adjacent seas or local evaporation account for the main precipitation during winter and early spring, while summer monsoon brings huge amounts of moisture from remote seas associated with higher temperature and larger precipitation amounts. Seasonal variations of the precipitation D-excess provide more details for changes in moisture sources. Higher D-excess values during winter and early spring are estimated to correspond to a lesser proportion of remote moisture, while lower D-excess values during summer and autumn correspond to larger remote moisture transported by summer monsoons. This generally agrees with the results of model analysis on single isobaric backward trajectories for air parcels during specific time periods. Results of this study imply that precipitation δ 18O and δD, as well as some related paleoclimate proxies such as δ 18O in speleothem and tree ring, and δD in plant-derived organic compounds and tree ring, currently cannot indicate changes in temperature or precipitation amount separately, but should be comprehensive proxies for monsoon climate. [ABSTRACT FROM AUTHOR]