Global vegetation response to extreme climate from 2001 to 2020
Exploring the spatiotemporal variations and response characteristics of global vegetation and extreme cli-mate is of great significance for addressing global climate change and improving ecosystem stability.Based on ERA5 climate data from the European Centre for Medium-Range Weather Forecasts and MODIS normalized difference vegetation index(NDVI)data,we used Sen's trend analysis,correlation analysis,and random forest regression model to explore the responses of NDVI of five vegetation types(boreal and temperate forest,tropical forest,other woody vegetation,grassland,and cropland)to 23 extreme climate indices from 2001 to 2020.The results showed that global NDVI showed an overall increasing trend from 2001 to 2020.The areas with the most significant growth trend was boreal and temperate forest,and the least significant growth trend occurred in cropland.In terms of extreme climate index,except for a few extreme high temperature and low temperature indices,the other indices showed an increasing trend.Across different vegetation areas,the extreme climate index that had the greatest influ-ence on NDVI was different.The results of correlation analysis showed that the indices with the greatest impact on NDVI in the boreal and temperate forest,tropical forest,other woody vegetation,grassland,and cropland were cold days,ice days,annual total precipitation,annual total precipitation,and annual total precipitation,respectively.The results of random forest analysis showed that the indices with the greatest impact on NDVI in each vegetation zone were cold days,warm night days,frost days,warm days,and the cold spell duration index,respectively.The reason for the different results between the two methods was that correlation analysis only reflected linear relation-ships between variables,while the random forest regression model could capture more complex nonlinear relation-ships.Our results showed that the response of global vegetation to extreme climate had significant regional differ-ences and complexities,which may result from interactions between different climate factors.
万方数据
