Changes in sessile oak (Quercus petraea) productivity under climate change by improved leaf phenology in the 3-PG model

Abstract

With global warming, the growing season is expected to increase for many regions of the Northern Hemisphere. It is therefore important to represent this mechanism in process-based forest growth models used in climate change impact analysis. 3-PG (Physiological Principles Predicting Growth) is a widely used process-based model that operates on stand-scale and monthly time steps. Yet, this model is currently not able to account for changes in the growing season of deciduous tree species. Therefore, we developed a method to model a dynamic growing season length with 3-PG. This method presents a novel approach to model dynamic leaf phenology with a monthly-time step model. We used a linear regression for calculating leaf onset and leaf senescence dates in response to temperature averages. The changes in the order of magnitude of days are then translated into monthly fractions. Further, we made some improvements to previous approaches of modelling the deciduous leaf cycle with 3-PG and we parameterized two model versions of 3-PG for sessile oak (Quercus petraea (Matt.) Liebl.) stands in Southwest Germany. The two model versions differed in presenting a i) constant and ii) dynamic growing season length. We used these model versions to estimate changes in sessile oak forest productivity under future climate scenarios. By comparing both model versions, we could disentangle the net effect of lengthening of the growing season on stem growth. By the end of the century, we observed on average 3 - 8 % increase in accumulated stem growth resulting from a lengthening of the growing season. At 1°C temperature increase, the lengthening of the growing season accounted for +3.3 % of accumulated stem growth (assuming unchanged precipitation). Leaf-unfolding advanced on average by 9 - 15 days and leaf senescence was delayed by 6 – 11 days for the period 2070-2100 compared to 1985-2015; for simulations of rcp 4.5 and 8.5 respectively. Overall, sessile oak productivity remained rather unchanged under future climate scenarios for Southwest Germany. Yet, we observed significant differences between sites (-6 % to +14 % in mean annual volume increment) as well as for different climate change scenarios and models (-2 % to +11 % in mean annual volume increment). We also modelled and discussed how assumptions on CO2 fertilization effects influenced 3-PG simulations.