Climate Research Domain BERTs: Pretraining, Adaptation, and Evaluation

Motivated by the pressing issue of climate change and the growing volume of data, we pretrain three new language models using climate change research papers published in top-tier journals. Adaptation of existing domain-specific models is utilized for CliSciBERT and SciClimateBERT and pretraining from scratch resulted in CliReBERT (Climate Research BERT). The performance assessment is performed on the climate change NLP benchmark ClimaBench. We evaluate SciBERT, ClimateBERT, BERT, RoBERTa and DistilRoBERTa - along with our new models - CliReBERT, CliSciBERT and SciClimateBERT - using five different random seeds on all seven ClimaBench datasets. CliReBERT achieves the highest overall performance with a macro-averaged F1 score of 65.45%, and outperforms all other models on three of the seven tasks. Additionally, CliReBERT demonstrates the most stable fine-tuning behavior, yielding the lowest average standard deviation across seeds (0.0118). The 5-fold stratified cross-validation on the SciDCC dataset showed that CliReBERT achieved the highest overall macro-average F1 score (53.75%), slightly outperforming RoBERTa and DistilRoBERTa, while the domain-adapted models underperformed their base counterparts. The superior performance of CliReBERT is accompanied by the lowest tokenizer fertility, suggesting appropriateness to model domain-specific vocabulary.