https://doi.org/10.1029/2023JG007839

Bowling, D. R., Schädel, C., Smith, K. R., Richardson, A. D., Bahn, M., Arain, M. A., Varlagin, A., Ouimette, A. P., Frank, J. M., Barr, A. G., Mammarella, I., Šigut, L., Foord, V., Burns, S. P., Montagnani, L., Litvak, M. E., Munger, J. W., Ikawa, H., Hollinger, D. Y., Blanken, P. D., Ueyama, M., Matteucci, G., Bernhofer, C., Bohrer, G., Iwata, H., Ibrom, A., Pilegaard, K., Spittlehouse, D. L., Kobayashi, H., Desai, A. R., Staebler, R. M., Black, T. A. (2024) Phenology of Photosynthesis in Winter-Dormant Temperate and Boreal Forests: Long-Term Observations From Flux Towers and Quantitative Evaluation of Phenology Models, John Wiley & Sons, Ltd, Journal of Geophysical Research: Biogeosciences, Vol 129, Iss 5, e2023JG007839, 2169-8953, https://doi.org/10.1029/2023JG007839

We examined the seasonality of photosynthesis in 46 evergreen needleleaf (evergreen needleleaf forests (ENF)) and deciduous broadleaf (deciduous broadleaf forests (DBF)) forests across North America and Eurasia. We quantified the onset and end (StartGPP and EndGPP) of photosynthesis in spring and autumn based on the response of net ecosystem exchange of CO2 to sunlight. To test the hypothesis that snowmelt is required for photosynthesis to begin, these were compared with end of snowmelt derived from soil temperature. ENF forests achieved 10% of summer photosynthetic capacity ~3 weeks before end of snowmelt, while DBF forests achieved that capacity ~4 weeks afterward. DBF forests increased photosynthetic capacity in spring faster (1.95% d-1) than ENF (1.10% d-1), and their active season length (EndGPP-StartGPP) was ~50 days shorter. We hypothesized that warming has influenced timing of the photosynthesis season. We found minimal evidence for long-term change in StartGPP, EndGPP, or air temperature, but their interannual anomalies were significantly correlated. Warmer weather was associated with earlier StartGPP (1.3-2.5 days °C-1) or later EndGPP (1.5-1.8 days °C-1, depending on forest type and month). Finally, we tested whether existing phenological models could predict StartGPP and EndGPP. For ENF forests, air temperature- and daylength-based models provided best predictions for StartGPP, while a chilling-degree-day model was best for EndGPP. The root mean square errors (RMSE) between predicted and observed StartGPP and EndGPP were 11.7 and 11.3 days, respectively. For DBF forests, temperature- and daylength-based models yielded the best results (RMSE 6.3 and 10.5 days)

keywords: photosynthesis, gross primary productivity, phenology, snowpack, forest, spring