Spatial and temporal patterns of vegetation- climate interactions

Abstract

In this thesis, I tried to understand the patterns and processes of vegetation trends at different spatial scales and resolution. The spatial variability in vegetation trends was related to inter and intra-annual changes in climate, mainly, temperature and precipitation. In this thesis, I applied non-parameteric linear regression (Sen’s slope) to calculate trends in vegetation and climatic variables.

The relationship between long-term vegetation trends and climate shifted between seasons in Trans-Himalaya. Phenological response to climate change can determine the strength and direction of feedback between the biosphere and the earth’s climate system. I evaluated the influence of changes in temperature and precipitation on vegetation phenology in the Central Asian highlands. The Central Asian highlands consists of arid and tundra climatic regions. Arid regions experience higher evapo-transpirational stress compared to tundra regions. Therefore, vegetation response to changes in climate can differ between arid and tundra regions. The changes in phenology was calculated by analyzing the changes in shape of uni-modal curve, which represents the vegetation annual biomass. Long-term changes in peak, skewness and kurtosis of the uni-modal curve correspond to peak biomass, timing of the peak biomass and length of the vegetation period respectively. Overall, arid regions showed higher changes in plant phenology compared to tundra with higher increase in peak biomass, earlier peaking and shorter vegetation period. Arid regions experienced higher temperature rise, but tundra regions experienced increase in snow and rain. The vegetation-climate relationship was evaluated using multiple linear regression and Cannonical Correlation Analysis. Both these analysis showed that the temperature and precipitation had an opposite effect on vegetation phenology. Warming conditions were related to earlier peaking and shortening the length of the vegetation period; while increase in precipitation was related to delayed peaking and extending the length of the vegetation period. Hence the additive and compensatory effect of temperature and precipitation determines the net-changes in vegetation phenology. However, vegetation phenology in tundra region is more sensitive to changes in rain whereas arid regions are sensitive to changes in snow and also show higher sensitivity to warming. Our results show that the sensitivity of phenological changes to temperature and precipitation form (snow/ rain) differs between climatic regions.

After evaluating patterns in vegetation trends at regional and biome level, a key question is to find a generic link between the processes which determine the plant biomass and observed vegetation patterns. The changes in net-photosynthesis determines the overall changes in the plant net primary productivity. Net-photosynthesis is sensitive to changes in temperature and shows a uni-modal response with leaf temperature. Hence, the value of T opt will determine the number of suitable/ unsuitable days for photosynthesis and trends in suitable/ unsuitable days determine the long-term changes in vegetation biomass. I conducted a meta-analysis of photosynthesis temperature response and evaluated the variation of net-photosynthesis with leaf temperatures between temperate and tropical plant species. The Harrell–Davis estimator showed that the net-photosynthesis started to decline after 32°C for both temperate and tropical species. I predicted the vegetation greening/ browning/ no-changing by evaluating trends in suitable/ unsuitable days for photosynthesis with a global T opt of 32°C. The predicted trends were matched with the observed trends evaluated using GIMMS NDVI data. This analysis was conducted first for global protected areas, croplands and pasturelands which showed a match of 51%, 48.1% and 48.5% respectively. The Bayesian probability of match of greening and browning trends between the predicted and observed was higher in protected areas and pasturelands than what could have been achieved as a random chance. As plants can thermo-regulate leaf temperature, changes in water stress can cause simultaneous changes in the net-photosynthesis. Hence we predicted vegetation trends based on long-term changes in water stress (SPEI). Change in temperature (80 million km 2 ) was more successful in explaining vegetation trends than simultaneous change in water-stress (72 million km 2 ) over the same period. This decline in net-photosynthesis above 32°C can be attributed to the thermal sensitivity of the photosynthesis enzymes mainly, Rubisco activase which starts denaturing above 30°C. This study shows that key physiological link between the processes occurring at a leaf level which can potentially determine the global terrestrial productivity.