Evaluation of the correlations of plant indices with atmospheric and biological variables using the Google Earth Engine

Natural vegetation is a vital component of ecosystem change models and conservation measures. The current study investigates the spatio-temporal dynamics of Normalized Differential Vegetation Index (NDVI) and Improved Vegetation Index (EVI) with atmospheric indices such as precipitation (P) and Multivariate Standardized Precipitation Index (MSPI); biological index such as evapotranspiration (ET) and soil moisture (SM); and soil index, land surface temperature parameter (LST) with the help of satellite images in Google Earth Engine system during 2000-2021 in Hormozgan province.

Google Earth Engine system images were used to evaluate plant indices with climatic and biological variables. NDVI and EVI indices, evaporation and transpiration and earth surface temperature from 16-day Terra satellite images and soil surface moisture (0-10 cm) from Open Land Map, during the period from 2000-2021 were extracted monthly by programming functions for the region. Rainfall and MSPI drought index were prepared with the help of 22-year statistics of the stations. A precipitation map was prepared using the Kriging geostatistics method in ArcGIS software. Using the PCA method, the drought index was calculated in different places with SPSS software. In order to better understand the changes in time, their anomaly plant indices were extracted over two decades. Correlation and change trend of parameters were calculated by Spearman and Mann-Kendall method.

Spearman's correlation coefficient and Man-Kendall's trending method between NDVI and EVI indices are significant at the 99% level of 0.84 and 0.67, respectively. The time distribution of plant indices showed that their annual and monthly changes are consistent with the amount of autumn cultivation. According to the hot and dry climate of the province, most of the rains are from January to March. The indices had the highest values during the months of January, March and April. This period corresponds to the season of planting and harvesting agricultural products, and the plants have reached their highest growth stage. Results showed the spatial changes of the plant indices correspond to the high rainfall areas and the edges of rivers and agricultural and garden lands. The appropriate overlap of precipitation zones and MSPI drought index with the spatial distribution of plant indices, ground surface temperature, and soil moisture indicates the determining role of precipitation in arid and semi-arid regions. The anomaly results showed that there is a weak positive correlation during the period of 2011-2000 due to frequent and severe droughts. In the second decade of 2011-2021, due to the decrease in the severity of the drought, the indicators showed a significantly strong correlation. According to the MSPI drought index, the western, central and eastern regions with the highest amounts of drought and lack of rainfall have the lowest amount of plant indices.

The expansion of the area has caused more concentration of vegetation indicators in some areas. The highest values of NDVI and EVI are in the area along the banks of rivers and water sources. These areas correspond to agricultural centers in the plains of Rodan, Minab, Shamil and Takht and parts of Haji Abad, which are geographically in the north-eastern regions and parts of the north. Vegetation dynamic trends and their relationship with other factors should be studied in the long term to establish a pattern. This study has been done as much as possible and according to the availability of a data set. Long-term studies can be done by extracting experimental variables at the regional level and in order to influence the population growth on the region. The findings of this study can help future studies related to vegetation and its relationship with other factors, nature protection and resource allocation.