Spring phenology in photosynthesis control and modeling for a temperate bog

Hongxing He ^1,2* Tim Moore ² Peter Lafleur ³ Oliver Sonnentag ⁴ Elyn R Humphreys ⁵ Mousong Wu ⁶ Nigel Roulet ²

• ¹ Great Lakes Forestry Centre, Natural Resources Canada, Sault Ste. Marie, Canada
• ² Department of Geography, Faculty of Science, McGill University, Montreal, Quebec, Canada
• ³ School of the Environment, Trent University, Peterborough, Ontario, Canada
• ⁴ Department of Geography, University of Montréal, Montréal, Canada
• ⁵ Department of Geography and Environmental Studies, Faculty of Arts and Social Sciences, Carleton University, Ottawa, Ontario, Canada
• ⁶ International Institute for Earth System Science, Nanjing University, Nanjing, Jiangsu Province, China

Predicting carbon dynamics of northern peatlands requires adequately representing vegetation phenology in terrestrial biosphere models. In this study, we analyzed the relative importance of various environmental controls to explain the start of the growing season by photosynthetic CO2 uptake for a temperate continental bog, using a multiple-year measured dataset comprising eddy covariance (EC), supporting environmental measurements and a digital image archive obtained with repeat photography. Vegetation in the studied bog is dominated by ‘evergreen’ shrubs and mosses. The vegetation phenological indices data including EVI, NDVI and green chromatic coordinate showed high correlations with ecosystem gross primary production (GPP) obtained from EC measurements, near-surface soil temperature, and the growing degree-day sum (∑GDD). We developed a new phenology scheme in the process-based CoupModel based on ∑GDD to represent the gradual greening of the 'evergreen' shrubs that regulate spring photosynthesis turn-on and increase. The new model simulated the earlier photosynthesis turn-on of the mosses and the shrubs' photosynthesis onset from a ∑GDD of 50 0C days. Incorporating a spring phenology subroutine of two canopies in the CoupModel improved model fit with the daily EC-derived GPP. Our results show without factoring in the spring phenology, CoupModel overestimates GPP by 24% and MODIS GPP had a 45% overestimation at the end of the spring season. Results from this study contribute to advancing our understanding of ecosystem dynamics and provide a foundation for refining ecosystem models to better capture the intricate interplay between phenology, carbon dynamics, and environmental conditions.