Evaluation of half-hour GPM-GSMAP_G (V04) and one-hour GPM-IMERG_F (V06) satellite precipitation products in East Azerbaijan Province

ABSTRACT:

Background and purpose:

Indirect measurement of precipitation through remote sensing is a practical way to achieve a comprehensive estimate of precipitation and to better understand the phenomenon of precipitation and its effective parameters in a wide spatial area with high spatial and temporal resolution. On the other hand, in satellite precipitation products, due to the use of different algorithms and correction methods by using ground rain gauge observations, there is a need to check the accuracy of these satellite precipitation products in different parts of the globe. Two satellite precipitation products with high and accurate temporal and spatial resolution in the new era are: 1-GPM-IMERGE precipitation product with a spatial resolution of 0.1 degrees and a temporal resolution of 0.5 hours, which is presented in three versions. 2- GPM-GSMAP rainfall product with a spatial resolution of 0.1 degrees and a time resolution of 1 hour, which this satellite rainfall product often provides in three versions. Most of the researches comparing the GSMAP and IMERGE products have compared these two products on a daily and larger time scale, and there are rare studies on an hourly scale. The purpose of this study is two research satellite precipitation products GPM-IMERG_F and GPM-GSMaP_G on a 6-hour scale (sub-daily) in East Azarbaijan province. Considering the lack of a dense rain gauge network in the region, along with the environmental disaster of drought in Lake Urmia, and the importance of rainfall information as a primary data in most water studies, it is necessary to move towards the use of satellite data to estimate rainfall in different areas; Therefore, it is necessary to evaluate the accuracy of these satellite precipitation products in the study area.

Materials and methods:

In this research, to evaluate the detection capability of satellite precipitation products, seven binary matching criteria including: probability of detection (POD), critical success index (CSI), false alarm ratio (FAR), Hick skill score index (HSS), bias frequency index (FBI), The correct ratio (PC) and the statistical index of true skill (TSS) and for the quantitative analysis of the accuracy of satellite precipitation products from six statistical indices including regression coefficient (R), root mean square error (RMSE), three bias indices (MBias, RBias, Bias) and the mean absolute error (MAE) index is used. Also, from the analysis of the Taylor diagram and the comparison of the spatial patterns of precipitation and the probability density curve of cumulative 6-hour precipitation and the performance of satellite precipitation products in terms of topography and altitude have been compared.

Results:

Despite the more accurate spatial and temporal resolution of both precipitation products, they still present a significant bias in some stations. Based on the Taylor diagram analysis, in all stations, the point corresponding to the IMERG_F satellite was closer to the observed point, and as a result, the IMERG_F product is better than the GSMaP_G product. Although both products had relatively similar spatial patterns in terms of statistical and binary indices, the GPM-IMERG_F product had much better accuracy and detection capability than the GPM-GSMaP_G product, so that the GPM-IMERG_F product has a cumulative probability density curve very close to the ground synoptic stations in terms of topography and altitude.

Conclusion:

GSMaP_G product has the best performance in the region. Compared to station observations, the IMERG_F product is underestimated, while the GSMaP_G product is overestimated. Compared to GSMaP_G, the IMERG_F product can better reproduce the 6-hour rainfall intensity PDF. In the region, the highest frequency of 6-hour rainfall occurs in the range of 0-0.1 mm and 5-10 mm in 6 hours, and the lowest frequency of rainfall occurs in the rainfall intensity of more than 20 mm in 6 hours. The IMERG_F product agrees well with the station observations in terms of frequency, when the rainfall intensity is greater than 1 mm in 6 hours. This study will be valuable to algorithm developers of these products as well as users of these products and can help in applications such as natural disaster risk reduction and hydrological modeling, especially in areas with a sparse rain gauge network.