Evaluation of process based model 3-PG for simulation of net primary production of Picea abies in northern and southern regions of Sweden under climate change

Detta är en Master-uppsats från SLU/Southern Swedish Forest Research Centre

Sammanfattning: The results of this evaluation reveal good performance of 3-PG and seems reasonable to use it for simulation of net primary production in Sweden. Subsequently,the NPP of Picea abies was simulated using 3-PG for 110 years in northern and southern Sweden under climate change. RCA3 generated climate data on two emission scenarios (A2 and B2) was used in the simulations as driving variables. The initial stand data and site factors were taken from well known sites in northern and southern Sweden to determine fertility rating input factor of 3-PG and to use for input data for Heureka StandWise and 3-PG for simulation and valida-tion. The outcome from the simulation of 2071-2100 in A2 and B2 scenario were summa-rized for 2071-75, 2076-80, 2081-85, 2086-90, 2091-95 & 2096-2100 and compared against their corresponding reference years (1961-1990). The average relative increment of NPP af-ter 110 years was 89,7% and 60,5 % for A2 and B2 in northern and 88,6% & 60,3% for A2 & B2 of southern Sweden respectively. Higher relative increase of temperature in autumn, spring & winter in northern Sweden led to higher relative increase of NPP in northern than Southern Sweden in both scenarios. Sensitivity testing of the model based on predicted NPP was carried out independently for temperature, rainfall and fertility rating. The result point-ed-out that NPP from 3-PG was more sensitive for fertility rating than for temperature and rainfall. Rainfall was almost indifferent for the test. Sensitivity of the factors considered in the exercise was found to be site dependent. Total biomass outputs from 3-PG and Heureka StandWise simulations were compared for validation. There was no significance difference between total biomass from the two models. Modeling efficiency was 78,5 % for northern and 89 % for southern Sweden. The average model bias explained the error with 8,6% and -3,2%; the mean absolute difference outcome was about 8,6% and 7% and the root mean square error was13% and 9,5% in northern & southern regions respectively. Overall, the results from this work suggest that there were possibilities to use 3-PG for predicting NPP in Sweden with due considerations of thinning operation, determination of fertility rating and Leaf area index outcomes.

  HÄR KAN DU HÄMTA UPPSATSEN I FULLTEXT. (följ länken till nästa sida)