Spatiotemporal Variation Of Vegetation Net Primary Productivity And Its Response To Climate Change In The Mongolian Plateau

Vegetation net primary productivity(NPP)is defined as the net accumulation of organic matter through photosynthesis by green plants per unit of time and space,a difference between gross primary productivity and autotrophic respiration.NPP represents the net amount of new carbon absorbed by green vegetation through photosynthesis,not only indicating the key parameter of vegetation activity,but also determining the ecosystem carbon sink/source and ecological process.Quantitative estimates of NPP variability and its response to climate change at regional to global scales are therefore significant for understanding changes in ecosystem quality and function,adjusting the ecological process and function,predicting terrestrial carbon cycle trends.Most recent satellite-based studies from different observation periods indicated that the vegetation productivity in the northern middle and high latitudes showed a great variability.However,due to the very complex characteristic of terrestrial ecosystems,different impacts of climate change and human activities,and natural hazard intensities,the changing trend and magnitude of vegetation variations/NPP at regional scale and the mechanisms behind that seem to be regionally dependent and still remain considerable uncertainties and controversies.Therefore,there still exists a great requirement to make clear,in great detail,the vegetation productivity variations at different regional scale for the fully understanding of global vegetation productivity changes.The Mongolian plateau,located at the central part of Eurasia,with an average elevation of 1580 m decreasing from west to east,is a very sensitive and vulnerable to climate change.The climate is characterized by continental arid and semiarid climate,and most part of the plateau was covered typical and desert steppes.Regardless of the geographical location,climate change,ecosystem types,and human activities,the Mongolian Plateau is a very unique and important ecological zone in the world due to the plateau is shared by two countries(China and Mongolia),hence,there is great need to make clear the spatial and temporal variation of vegetation NPP in the plateau,particularly under background of the present global climate change.This would be valuable to enrich the global change studies,and verify the vegetation response characteristic to climate change in different regions of the Eurasia.Furthermore,we could identify the differences in vegetation responses to climate change between the Mongolia and Inner Mongolia under the condition of the different land use and pasturing economic activities.In this study,the spatial and temporal variation of NPP and vegetation phenology and their response to climate change in the plateau from 1982 to 2011 is thoroughly examined by employing CASA ecosystem model and three phenology extraction methods(Inflection points from the Logistic curvature of cumulative NDVI?Rate of change of Logistic curvature and rate of change of NDVI).The time series of second and third generation Global Inventory Modeling and Mapping Studies Normalized Difference Vegetation Index(GIMMS NDVI),the Moderate Resolution Imaging Spectroradiometer(MODIS)NDVI,climatic data from 108 meteorological stations and 22 solar radiation stations,vegetation type data and ground observed NPP were used for the study.The phenology impacts in NPP variation is also further identified in the study.The main study results can be concluded:(1)In the light of the facts that the values of maximum light use efficiency in the current CASA model is not separated for meadow steppe,typical steppe and desert steppe,we modified the different maximum light use efficiency values for different steppes based on the accuracy minimal principle between field observed NPP and CASA modeled NPP,and then these different maximum light use efficiency values were used to model the NPP variations in the Mongolian Plateau during 1982-2011.The results indicated that the estimated maximum light use efficiency values are 0.654gC·MJ-1,0.553 g C·MJ-1 and 0.511 g C·MJ-1 for meadow steppe,typical steppe and desert steppe,respectively.Compared to the result without modification of maximum light use efficiency(0.541 g C·MJ-1),the determination coefficient and relative means square error(RMSE)between modeled and observed NPP was improved 0.024 and 2.62 g C/m2 when maximum light use efficiency was modified.This could imply that modifications for more detailed vegetation types such species composition could further improve the NPP modeling accuracy when using CASA model.(2)Due to the influences of the vapor from Arctic Ocean moisture in the north(i.e.northern Mongolia)and the vapor of the Pacific moisture in the northeast(i.e.northeastern Inner Mongolia),climatic zone in the plateau is characterized by the north-south(northeast-southwest)gradient from cold humid and semi-humid climate in the northern mountain regions to temperate semiarid and arid climate in central and southeast parts.Influenced by the climate,the vegetation transected forest,forest meadow,typical steppe,desert steppe,Gobi-desert from north to south and Gobi-desert,desert steppe,typical steppe,farming-pasturing ecozone,forest meadow and forest.Accordingly,the highest NPP(more than 400 g C/m2/year,highest:711.29 g C/m2/year)can be observed in the forest in Da Xingan mountains in Inner Mongolia and Sayan Mountains,and Hentiy Mountain regions in Mongolia where precipitation ranged from 300-450 mm and 250-350 mm,respectively,and which gradually decreased from north to south and northeast to southwest.The lowest values can be observed in the southwestern parts of the plateau,with a less than 100 g C/m2/year,particularly in the Gobi desert regions,where NPP decreased to below 20 g C/m2/year.(3)NPP in the Mongolian Plateau fluctuated around 240 g C/m2/year and no significant trend can be observed from 1982 to 2011.However,the piecewise linear regression model predicted a turning point(TP)of 1993 with increasing trend before 1993 and decreasing trend after 1993 to2011.TP in the plateau seems to 4 years earlier than Eurasian continent and the decreasing trend from 1993 was slower than Eurasia.The climate record indicates that the response of NPP variability to climate change is not completely same with Eurasian vegetation.Precipitation is decisive climatic factor causing NPP variability in the plateau,and there are almost same fluctuation between NPP and precipitation and completely same TP at the entire plateau scale.The correlation analysis further confirmed this result,with a correlation coefficient of 0.88.However,there are strong spatial heterogeneity in the NPP variation and its response to climate change,TP.On the whole,the NPP variation in the northern mountain regions such Hovsgol mountains?Sayan mountains?Hentiy mountains in Mongolia and Da Xingan mountain forest regions in Inner Mongolia is generally governed by temperature and NPP in central and southwestern steppe grasslands regions is controlled by precipitation.The climate record shows that drying and warming trend is very significant in the plateau,particularly since 1990s,and shortage of water resource highlighted a prominent environmental issue in the plateau and main reasons of TP of the NPP.(4)Influenced by the differences in seasonal and monthly climate change,there exists strong differences in seasonal NPP variations and their response to seasonal climate change.During the entire observation period of 1982-2011,only spring NPP experienced weak increasing trend,while summer and autumn NPP keep more stability or weak decreasing trend insignificantly.The TP of each season and two climatic variables did not correspond each other,indicating that the reasons of the seasonal NPP TP are very complex and difficult to explain it by TP of climatic variables.Summer precipitation has strong beneficial to vegetation growth in the plateau,followed by the spring and autumn.There are no significant correlation seasonal NPP and temperature,indicating that there did not exist strong response of NPP to temperature during previous 30 years.However,spring and autumn temperature has positive impact on two seasons NPP,while summer temperature has negative impact on summer NPP.At the plateau scale,summer precipitation produced one season(three months)lag-time effect on autumn NPP,while there did not exist other lag-time effect of temperature and precipitation on NPP variation.Due to strong spatial heterogeneities in distribution of both climate and vegetation types,there are apparent spatial heterogeneities in responses of seasonal and monthly NPP variations to climate change,including lag time responses,particularly,between the forest regions in the northern Mongolia and northeastern Inner Mongolia and steppes in central and southwestern part of the plateau.(5)Comparison between observed phenology and modeled phenology metrics indicated that the average phenology of inflection points from the Logistic curvature of cumulative NDVI and rate of change of NDVI performed acceptable results in phenology modeling.The correlation coefficient and RMSE between observed and modeled values are 0.634(SOS),0.277(EOS),5.3 days(SOS)and 5.8 days(EOS)for start of growing season(SOS)and end of growing season(EOS),respectively,a higher value than mono phenology model.(6)During the 30-year observation period(1982-2011),no significant advancement or delay can be observed in SOS in the Mongolian Plateau,however,SOS firstly advanced from 1982 to 2001 and then delayed after 2001 until now.Although EOS experienced a weak advancement for the period of 1982-2011,there are still two distinct periods with an opposite trends before and after 1995,with a delay before 1995 and advancement after 1995.Influenced by the weak advancement of EOS,LOS in the study area showed a weak shortening trend for previous 30 years,however,LOS first lengthened before 1998 and reversed afterward.The advancement of SOS,delay of EOS and lengthened LOS before the late 1990s is generally consistent with that observed in northern mid-and high latitudes during the same period.The spatial distribution analysis indicated that northern humid and sub-humid regions in the plateau have earlier SOS,later EOS and longer LOS,while the regions characterized by arid and semiarid climate have later SOS,earlier EOS and shorter LOS.The correlation analysis indicated that precipitation in April and temperature in May played important role in advancement of SOS due to the lag time effect of precipitation.Precipitation in July and August played a very significant role in delay of EOS and stronger than that in September and October.Compared to precipitation,there was no significant correlation between temperature in September and October and EOS,suggesting that it is difficult to explain the variation of EOS using monthly temperature.The phenology metrics of different vegetation types experienced different trends and response characteristics to climate change due to their different growing environment and biological features.(7)The correlation analysis between three phenological parameters and NPP indicate that the advancement(delay)of the SOS could increase(decrease)the NPP in April and May,but would not necessarily have a significant impacts on growing season NPP due to the impact of summer droughts.However,the delay(advancement)of EOS not only increase(decrease)the NPP in September and October,but also increase(decrease)growing season NPP most probably due to larger proportion of autumn NPP than spring NPP in growing season total NPP.The extension of LOS could significantly increase the growing season NPP(R=0.54,P=0.002),with an increment of 2.69 g C/m2/day.It can be clearly seen that the impact of phenological changes in NPP varied by regions,month,season and vegetation types.The major differences mostly occurred between the forest regions in the north and northeastern in the plateau and steppe grassland distributed regions in the central and southwestern parts of the plateau.The summer precipitation and its induced summer NPP have strong impacts on annual growing season NPP variation,and stronger than SOS advanced and EOS delayed NPP increases,mostly due to the larger proportion of summer NPP in growing season total NPP.In this paper,the maximum light use efficiency was optimized for meadow,typical and desert steppe in the Mongolian Plateau based on the principle of minimum error using CASA model,field observed NPP and grassland distribution data.The modeling accuracy of NPP was improved using different maximum light use efficiency for different steppe types.Based on this,the spatiotemporal dynamics of NPP and Its response to climate change during 1982-2011 in the plateau was thoroughly examined.The complementary advantages of three phenology extraction methods(Inflection points from the Logistic curvature of cumulative NDVI?Rate of change of Logistic curvature and rate of change of NDVI)was further tested in the plateau and modeling accuracy of the phenology was increased through computing mean value of phenological parameters.Based on this,we deeply analyzed the spatiotemporal changes of phenology and its response to climate change and phenological impacts on NPP in the plateau.