Modeling the Impact of Climate Change on Net Primary Productivity (NPP) of Selected Forest Ecosystems in Nilgiri Biosphere Reserve,India

Abstract

As an inevitable component of global terrestrial ecosystem, vegetation plays a crucial role in energy transfer, carbon cycle, water balance and climate regulation. Its response to environment change has been considered as one of the key fields of ecological research. Net primary productivity (NPP) is defined as the net amount of carbon taken in by plants via photosynthesis, and is equal to the difference between the carbon assimilated during photosynthesis and that released during plant respiration. The present study tried to investigate mainly the following events: i) Land use and land cover changes over the NBR, India over 18 years (2000 to 2018) using MODIS data wherein, we, estimated the LULC of NBR, identified the changes in LULC in 18 years, and checked the accuracy of the LULC of 2018 with ground truth data using kappa coefficient. In the study, the importance of monitoring of LULC changes is revealed by the decrement in areas of closed shrub lands and grasslands during the period 2000-2018. Also, the human activities within the buffer regions of the NBR were understood by the increment of areas of classes like croplands and cropland/natural vegetation mosaics during the study period; Secondly, (ii) Assessed the impact of climate change on NPP over NBR during the period (1981 to 2019): wherein, The NPP that were expected to realize in location as well as in different forested ecosystems of the region were estimated through satellite derived data using CASA model. The data taken for modeling study was for a period of 38 years i.e., from 2018 to 2019. It was observed that the mean NPP of NBR was ranging from 47.2 to 183.73 (g C m-2 month-1). NPP trend ranged from (-0.154 to 0.176) for the period. Seasonal NPP of post monsoon season (ON) showed an increasing trend of NPP highlighting the positive influence of precipitation on NPP. During the study period, the trends observed per year were 0.14 W m-2 decrease in solar radiation, 0.02 0C increase in Air temperature, 0.10 mm increase in Precipitation and 0.06 g C m-2 month-1 increase in NPP respectively. Moreover, the estimated NPP showed spatial variation across the region of different LULC. Very large NPP (>103 g C m -2 month -1) for Evergreen Broadleaf Forest together with its smaller counterpart (57 g C m -2 month -1) for Persistent Wetlands were estimated over study region among different LULC. Highest overall mean NPP was in the order EBF>CNVM > DBF=MF>WS; and finally we, predicted the present and future NPP (2018 to 2028) using SARIMAX (0, 0, 2) (1, 1, 0) time series model using the output of CASA model. We divided the data sets in to two parts viz., 1981 to 2018 and 2010 to 2018; first set to build the model (R 2 = 0.552) and the second set for validation of the fitted model. We took data 1981-2010 and forecasted the NPP for the rest Eight years taking the air temperature, precipitation and solar radiation (2010 - 2018) and compared with the observed values of NPP (CASA derived NPP from 2010 to 2018). As there was good correlation between the observed and the predicted (8 years) (r = 0.63), hence validated the model and used for future prediction. As there were no explanatory variables for prediction (Ten years), hence assumed that the same trend for each explanatory variable for next Ten years to make the forecast of NPP and hence fitted individual ARIMA model for each of the explanatory variables (Seasonal ARIMA) viz., solar radiation, air temperature and precipitation. Later on, We used the SARIMAX(0, 0, 2) (1, 1, 0) model for prediction using the forecasted individual explanatory variables for Ten years and finally derived equation for predicted NPP using estimate for lag values as mentioned in the model fit statistic of NPP based on SARIMAX (0, 0, 2) (1, 1, 0).