Vegetation Autumn Phenology And Its Responses To Environmental Changes In The Mid-high Latitudes Of The Northern Hemisphere Based On Remote Sensing Data

The interannual variation in vegetation phenology partly reflects the dynamic responses of terrestrial ecosystems to global environmental changes.Vegetation phenology is largely regulated by temperature,precipitation,radiation,and its own circadian rhythms.Not only is it highly sensitive to climate change,but also it affects the energy flow of the terrestrial ecosystems and plays an important role in the global cycles of carbon,nitrogen and water fluxes.Previous studies were usually centered on the start of the growing season(SOS),while there is less concentration on the end of the growing season(EOS).Under the background of global environmental changes,autumn phenology not only regulates the duration of annual vegetation growth,but also it affects the carbon budget of the whole ecosystem.Therefore,it is necessary to deepen the study of autumn phenology.Researchers suggest that warming,increased precipitation,and enhanced radiation are prone to delay autumn phenology.Besides climatic factors,the changes of atmospheric CO₂ and nitrogen content also have an important impact on plant growth via photosynthesis,but the existing research on autumn phenology seldom pays attention to atmospheric CO₂ and nitrogen changes.In addition,revealing the main controlling factors affecting autumn phenology is the key basis for developing autumn phenological models and predicting future phenological trends.But the relative contributions of climatic changes and atmospheric composition changes(CO₂ and nitrogen)to autumn phenology is still unknown.Therefore,to study the influential mechanisms of atmospheric composition changes(CO₂ and nitrogen)on plant phenology in autumn,and compare the relative effects of atmospheric composition and climatic change is helpful to enrich our understanding of autumn phenology,to build a more complete autumn phenology model,and to better predict the changing trend of autumn phenology under future climate change.In view of the above points,this thesis adopted the longest and widely used GIMMS NDVI3g data set(1982?2015)to quantitatively extract the end of growing season(EOS)in the mid-high latitudes(>30°N)of the northern hemisphere.Three phenological extraction methods were used here,namely,dynamic-threshold method,piecewise logistic function,and double logistic function.Then,the spatial characteristics,long-term interannual trends of EOS,and the differences between the results of extraction algorithms were analyzed.Afterwards,the partial correlation analysis and the partial least squares regression analysis were used to explore the impacts of precipitation,radiation,temperature,atmospheric CO₂,and nitrogen deposition on EOS.The dominant factors of EOS were determined at each pixel,and the contribution of each environmental factor to EOS was mainly analyzed from the perspective of plant communities.In addition,we also discussed the different impacts from both atmospheric and climate changes on EOS in detail.The main conclusions are as follows:(1)The EOS extracted by the three phenological extraction methods were largely consistent in the spatial distribution and temporal variation,but there were still differences in the specific date.Generally,the EOS extracted by the two fitting methods were earlier than that extracted by the dynamic-threshold method,and the overall error was within 15days.In order to reduce the difference between phenological extraction methods,this thesis used the average of the three methods as the final EOS.(2)The temporal changes of EOS in the mid-high latitudes of northern hemisphere was quantified,and the opposite trends were found in the mid-latitude and high-latitude regions.From 1982 to 2015,58.5%of the vegetated pixels showed a trend of advancement in EOS,of which 13.6%were significant(p<0.05),and the average advancing rate was 3.8±4.6 days per decade.The significant advancements in EOS were mainly found in Central Siberia and northwestern North America.The remaining pixels(41.5%)showed a delaying trend,of which 10.3%were significantly delayed,and the average delaying rate was 2.8±2.9 days per decade.The significant delaying trend mainly occurred in northwestern North America,central Europe,and East Asia.From the perspective of plant community,the delayed EOS was quite obvious in temperate broadleaf and mixed forests(TBMF),while it showed an overall advancing trend in tundra(TUN).The sensitivity of EOS to environmental factors was also significantly different across different biome types.Forest types were highly sensitive to temperature and radiation,grasslands were very sensitive to changes in precipitation,while TBMF and TUN showed strong sensitivity to atmospheric CO₂ and nitrogen deposition.(3)The relative contributions of climate change and atmospheric composition changes to autumn phenology was also quantified,and the result showed asymmetric responses of autumn phenology to climate change,atmospheric CO₂ and nitrogen deposition changes among biome types.In the mid-high latitudes of the northern hemisphere,climate change played a key role in the autumn phenological change at about70%of the regions,while the rest of the areas(30%)were mainly controlled by atmospheric CO₂ and nitrogen deposition.Under global warming,the influence of temperature on EOS forests and tundra was mostly positive for forests and tundra,while it was more negative for grasslands in arid and semi-arid areas.The increase in precipitation had a particularly significant effect on the delay of EOS in grasslands.In forests,the positive response of EOS to radiation occupied more areas than that of the negative response,while it was opposite for grasslands.The increase in nitrogen deposition alleviates the nitrogen limitation of the tundra,and it had a positive effect on the EOS at TUN.The negative impact of atmospheric CO₂ on EOS was more obvious in grasslands.These results are of vital importance to the parameterization of autumn phenological models and the prediction of global vegetation phenological dynamics.