Fine-tuning satellite-based rainfall estimates

4th International Symposium on LAPAN-IPB Satellite for Food Security and Environmental Monitoring

农业科学;生态环境科学

Harsa, Hastuadi^1 ; Buono, Agus^2 ; Hidayat, Rahmat^3 ; Achyar, Jaumil^4 ; Noviati, Sri^1 ; Kurniawan, Roni^1 ; Praja, Alfan S.^1

Research and Development Center, Indonesia Meteorology Climatology and Geophysics Agency, Jl. Angkasa I No. 2, Kemayoran, Jakarta

10720, Indonesia^1

Computer Science Department, Mathematics and Natural Science Faculty, Bogor Agricultural University, Bogor, Indonesia^2

Geophysics and Meteorology Department, Mathematics and Natural Science Faculty, Bogor Agricultural University, Bogor, Indonesia^3

Database Center, Indonesia Meteorology Climatology and Geophysics Agency, Jl. Angkasa I No. 2, Kemayoran, Jakarta

10720, Indonesia^4

关键词: Bias correction;    Bias-correction methods;    Ground observations;    Inverse distance weighting;    Mean and standard deviations;    Observation data;    Rainfall estimates;    Spatial composition;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/149/1/012047/pdf
DOI  :  10.1088/1755-1315/149/1/012047

来源: IOP

【 摘 要 】

Rainfall datasets are available from various sources, including satellite estimates and ground observation. The locations of ground observation scatter sparsely. Therefore, the use of satellite estimates is advantageous, because satellite estimates can provide data on places where the ground observations do not present. However, in general, the satellite estimates data contain bias, since they are product of algorithms that transform the sensors response into rainfall values. Another cause may come from the number of ground observations used by the algorithms as the reference in determining the rainfall values. This paper describe the application of bias correction method to modify the satellite-based dataset by adding a number of ground observation locations that have not been used before by the algorithm. The bias correction was performed by utilizing Quantile Mapping procedure between ground observation data and satellite estimates data. Since Quantile Mapping required mean and standard deviation of both the reference and the being-corrected data, thus the Inverse Distance Weighting scheme was applied beforehand to the mean and standard deviation of the observation data in order to provide a spatial composition of them, which were originally scattered. Therefore, it was possible to provide a reference data point at the same location with that of the satellite estimates. The results show that the new dataset have statistically better representation of the rainfall values recorded by the ground observation than the previous dataset.