CGIAR Research Program for Climate Change, Agriculture and Food Security (CCAFS), International Rice Research Institute (IRRI), IRRI-CCAFS Office, Agricultural Genetics Institute, Km 2 Pham Van Dong Ave, Tu Liem District, Hanoi, Vietnam.
International Research Institute for Climate and Society, The Earth Institute at Columbia University, New York, USA.
Sci Rep. 2022 Jan 11;12(1):485. doi: 10.1038/s41598-021-04380-8.
High-resolution reliable rainfall datasets are vital for agricultural, hydrological, and weather-related applications. The accuracy of satellite estimates has a significant effect on simulation models in particular crop simulation models, which are highly sensitive to rainfall amounts, distribution, and intensity. In this study, we evaluated five widely used operational satellite rainfall estimates: CHIRP, CHIRPS, CPC, CMORPH, and GSMaP. These products are evaluated by comparing with the latest improved Vietnam-gridded rainfall data to determine their suitability for use in impact assessment models. CHIRP/S products are significantly better than CMORPH, CPC, and GsMAP with higher skill, low bias, showing a high correlation coefficient with observed data, and low mean absolute error and root mean square error. The rainfall detection ability of these products shows that CHIRP outperforms the other products with a high probability of detection (POD) scores. The performance of the different rainfall datasets in simulating maize yields across Vietnam shows that VnGP and CHIRP/S were capable of producing good estimates of average maize yields with RMSE ranging from 536 kg/ha (VnGP), 715 kg/ha (CHIRPS), 737 kg/ha (CHIRP), 759 kg/ha (GsMAP), 878 kg/ha (CMORPH) to 949 kg/ha (CPC). We illustrated that there is a potential for use of satellite rainfall estimates to overcome the issues of data scarcity in regions with sparse rain gauges.
高分辨率可靠的降雨数据集对于农业、水文和与天气相关的应用至关重要。卫星估计的准确性对模拟模型，特别是对作物模拟模型，有重大影响，因为这些模型对降雨量、分布和强度高度敏感。在这项研究中，我们评估了五种广泛使用的业务卫星降雨估计产品：CHIRP、CHIRPS、CPC、CMORPH 和 GSMaP。通过将这些产品与最新的改进后的越南格网降雨数据进行比较，评估了它们在影响评估模型中的适用性。CHIRP/S 产品的表现明显优于 CMORPH、CPC 和 GsMAP，其技能更高、偏差更低，与观测数据具有高度相关性，且平均绝对误差和均方根误差较低。这些产品的降雨检测能力表明，CHIRP 的检测概率（POD）评分较高，表现优于其他产品。不同降雨数据集在模拟越南玉米产量方面的表现表明，VnGP 和 CHIRP/S 能够很好地估计平均玉米产量，其均方根误差范围为 536 公斤/公顷（VnGP）、715 公斤/公顷（CHIRPS）、737 公斤/公顷（CHIRP）、759 公斤/公顷（GsMAP）、878 公斤/公顷（CMORPH）和 949 公斤/公顷（CPC）。我们说明，卫星降雨估计有潜力克服降雨量稀少地区数据稀缺的问题。
