Easy-to-use spatial random-forest-based downscaling-calibration method for producing precipitation data with high resolution and high accuracy
نویسندگان
چکیده
Abstract. Precipitation data with high resolution and accuracy are significantly important in numerous hydrological applications. To enhance the spatial of satellite-based precipitation products, an easy-to-use downscaling-calibration method based on a random forest (SRF-DC) is proposed this study, where autocorrelation measurements between neighboring locations considered. SRF-DC consists two main stages. First, downscaled by SRF incorporation high-resolution variables including latitude, longitude, normalized difference vegetation index (NDVI), digital elevation model (DEM), terrain slope, aspect, relief land surface temperatures. Then, calibrated rain gauge observations aforementioned variables. The monthly Integrated MultisatellitE Retrievals for Global Measurement (IMERG) over Sichuan Province, China, from 2015 to 2019 was processed using SRF-DC, its results were compared those classical methods geographically weighted regression (GWR), artificial neural network (ANN), (RF), kriging interpolation only measurements, bilinear interpolation-based downscaling then SRF-based calibration (Bi-SRF), geographical analysis (GDA)-based (SRF-GDA). Comparative analyses respect root mean square error (RMSE), absolute (MAE) correlation coefficient (CC) demonstrate that (1) outperforms as well original IMERG; (2) estimation slightly more accurate than annually fraction disaggregation method; (3) perform better (Bi-SRF) GDA-based (SRF-GDA); (4) GWR ANN, whereas map loses detailed patterns; (5) variable-importance rank RF, interpolated most variable, indicating significance incorporating estimation.
منابع مشابه
Forest Classification Using High Spectral and Spatial Resolution Data
A modern airborne imaging spectrometer provides an exact observation of the object’s spectral properties. Besides, the spatial resolution of imaging spectrometers is good and instruments provide information about object texture. Many nature objects can be identified by their characteristic reflection. However, the spectral properties of certain plants are similar and the identification made onl...
متن کاملMeRIP-PF: An Easy-to-use Pipeline for High-resolution Peak-finding in MeRIP-Seq Data
RNA modifications, especially methylation of the N(6) position of adenosine (A)-m(6)A, represent an emerging research frontier in RNA biology. With the rapid development of high-throughput sequencing technology, in-depth study of m(6)A distribution and function relevance becomes feasible. However, a robust method to effectively identify m(6)A-modified regions has not been available yet. Here, w...
متن کاملA High-Reliability, High-Resolution Method for Land Cover Classification Into Forest and Non-forest
We present several methods for per-region land-cover classification based on distances on probability distributions and whole-region probabilities. We present results on using this method for locating forest areas in high-resolution aerial images with very high reliability, achieving more than 95% accuracy, using raw radiometric channels as well as derived color and texture features. Region bou...
متن کاملA New Structural Matching Method Based on Linear Features for High Resolution Satellite Images
Along with commercial accessibility of high resolution satellite images in recent decades, the issue of extracting accurate 3D spatial information in many fields became the centre of attention and applications related to photogrammetry and remote sensing has increased. To extract such information, the images should be geo-referenced. The procedure of georeferencing is done in four main steps...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Hydrology and Earth System Sciences
سال: 2021
ISSN: ['1607-7938', '1027-5606']
DOI: https://doi.org/10.5194/hess-25-5667-2021