Predicting spring phenology in deciduous broadleaf forests: NEON phenology forecasting community challenge
Kathryn I. Wheeler, Michael C. Dietze, David LeBauer, Jody A. Peters, Andrew D. Richardson, Arun Ross, R. Quinn Thomas, Kai Zhu, U. Narayan Bhat, Stephan B. Munch, Raphaela Floreani Buzbee, Min Chen, Benjamin R. Goldstein, Jessica Guo, Dalei Hao, Chris Jones, Mira Kelly-Fair, Haoran Liu, Charlotte Malmborg, Naresh Neupane, Debasmita Pal, Vaughn Shirey, Yiluan Song, McKalee Steen, Eric A. Vance, Whitney M. Woelmer, Jacob H. Wynne, Luke J. Zachmann
Abstract
Accurate models are important to predict how global climate change will continue to alter plant phenology and near-term ecological forecasts can be used to iteratively improve models and evaluate predictions that are made a priori. The Ecological Forecasting Initiative's National Ecological Observatory Network (NEON) Forecasting Challenge, is an open challenge to the community to forecast daily greenness values, measured through digital images collected by the PhenoCam Network at NEON sites before the data are collected. For the first round of the challenge, which is presented here, we forecasted canopy greenness throughout the spring at eight deciduous broadleaf sites to investigate when, where, and for what model type phenology forecast skill is highest. A total of 192,536 predictions were submitted, representing eighteen models, including a persistence and a day of year mean null models. We found that overall forecast skill was highest when forecasting earlier in the greenup curve compared to the end, for shorter lead times, for sites that greened up earlier, and when submitting forecasts during times other than near budburst. The models based on day of year historical mean had the highest predictive skill across the challenge period. In this first round of the challenge, by synthesizing across forecasts, we started to elucidate what factors affect the predictive skill of near-term phenology forecasts.