Agricultural drought in a typical semiarid Mediterranean environment is investigated during the growing seasons of 1997 through 2020 using a combination of optical and thermal sensors onboard Landsat satellites. The combination of the Normalized Difference Vegetation Index – Land Surface Temperature (NDVI- LST) space was able to distinguish between drought and non-drought years. A distinct trapezoidal shape was clearly defined during non-drought years, reflecting the strong negative correlation between NDVI and LST. The NDVI-LST space was poorly defined for drought-stricken years with no clear link between the two parameters. The non-universal relationship between LST and NDVI was addressed using the Monin- Obukhov similarity formulation which shows that the widely observed convergence of LST at high NDVI values could be explained by the asymptotic nature of LST against surface roughness length for non-stressed vegetation.

The NDVI-LST space was compared with seasonal and annual precipitation and different SPI windows to check the ability of the remote sensing metric to identify drought. A high correlation exists between the NDVI-LST space on the one hand and the 9- month, annual precipitation, the SPI-6, SPI-9 and SPI-12 windows, with correlation coefficients of 0.74, 0.76, 0.76, 0.80, and 0.80, respectively, which are statistically significant.

在1997年至2020年的生长季节中，使用Landsat卫星上的光学和热传感器的组合对典型的半干旱地中海环境中的农业干旱进行了调查。归一化植被指数–地表温度（NDVI-LST）空间的组合能够区分干旱年和非干旱年。在非干旱年份清楚地定义了明显的梯形形状，反映了NDVI和LST之间的强烈负相关。NDVI-LST空间在干旱年份定义不清，两个参数之间没有明确的联系。

将NDVI-LST空间与季节性和年度降水以及不同的SPI窗口进行了比较，以检查遥感指标识别干旱的能力。一方面NDVI-LST空间与9个月的年降水量，SPI-6，SPI-9和SPI-12窗口之间存在高度相关，相关系数分别为0.74、0.76、0.76、0.80和分别为0.80，具有统计学意义。