土壤质地改变对CLDAS/Noah-MP土壤湿度模拟的影响研究

Impact of Soil Texture on the Simulation of CLDAS/Noah-MP on Simulating Soil Moisture

陆面模式可以模拟获得高时空分辨率连续的多层土壤湿度,但其精度受到地表参数的影响,土壤质地就是其中之一,本文利用CLDAS-V2.0(中国气象局陆面数据同化系统)驱动Noah-MP模式,开展了其对土壤湿度模拟的影响研究。结果表明,利用中国第二次土壤调查数据制作的中国区土壤质地数据(SNSS)与模式自带土壤质地数据(FAO)模拟的2014年中国区域土壤湿度日均值存在显著差异,0~10cm深度23.2%的区域差异性大于10%,74.9%的区域差异性大于1%;10~40 cm深度20.8%的区域差异性大于10%,69.8%的区域差异性大于1%。从时间序列来看,两组实验模拟结果均能基本反映土壤湿度随时间变化的规律,SNSS在0~10 cm处模拟结果表现更好,但在10~40 cm处出现低估现象。从空间分布分析,使用SNSS土壤质地类型之后,CLDAS/Noah-MP土壤湿度模拟结果与观测值的偏差为负值的区域较多,尤其是10~40 cm深度,大多数区域模拟值均存在低估;与FAO模拟结果比较,SNSS在东南和西南地区0~10 cm和10~40 cm深度的模拟效果有所改进;东北地区0~10 cm深度SNSS的模拟效果好于FAO,但10~40 cm深度的模拟精度两者相差较小。

Soil moisture is a key physical quantity in the land-atmosphere interaction,and also an important parameter in the research of land surface model simulation. However,it is very difficult to obtain the spatiotemporal continuous data in high resolution. Land surface models(LSMs)can simulate the multilayer soil moisture with high spatial and temporal resolution,the community Noah land surface model with multiparameterization options(Noah-MP)has been proved to be a good model for simulating soil moisture in China,but the accuracy needs to be further improved. Soil texture data is an important input data of Noah-MP,and its quality affects the accuracy and reliability of simulation results,it is of great significance to study the effect of soil texture on Noah-MP on simulating soil moisture. Here,we used two different soil texture data to study the effect of soil texture on simulation of Noah-MP LSM driven by CLDAS-V2. 0. The results showed that there was a significant difference between the simulated daily mean of soil moisture by the soil texture data from the Second Soil survey of China(SNSS)and the mode inherent soil texture data(FAO)in 2014. The difference of the 0~10 cm depth simulation was greater than 10% in 23. 2% areas and 1% in 74. 9% areas. For 10~40 cm depth,they were 20. 8% and 69. 8% respectively. From the analysis of time series,the simulation results of the two group tests of SNSS and FAO can basically reflect the law of soil moisture changing with time. According to the daily results of correlation coefficient,bias and RMSE analysis between simulation and observation,the simulation results of SNSS at 0~10 cm depth were better than those of FAO in almost every day,but in 10~40 cm depth,the simulation results simulated by SNSS are obviously underestimated than the observations. From the analysis of spatial distribution,the soil moisture simulation results of CLDAS/Noah-MP showed that there were many areas with negative deviation from the observed values,especially the depth of 10~40 cm,and most of the simulated values were underestimated. Compared with FAO,simulation results simulated by SNSS in the southeast and southwest region at 0~10 cm depth and 10~40 cm depth were improved. The simulation simulated by SNSS at 0~10 cm depth was better than that by FAO in northeast China,but there was little difference between them at 10~40 cm depth. This study provides a scientific basis for the study and application of soil parameter data in CLDAS/Noah-MP.