Feature tuning improves MAXENT predictions of the potential distribution of Pedicularis longiflora Rudolph and its variant

Abstract

Pedicularis longiflora Rudolph and its variant (P. longiflora var. tubiformis (Klotzsch) Tsoong) are alpine plants and traditional Chinese medicines with important medicinal value, and future climate changes may have an adverse impact on their geographic distribution. The maximum entropy (MAXENT) model has the outstanding ability to predict the potential distribution region of species under climate change. Therefore, given the importance of the parameter settings of feature classes (FCs) and the regularization multiplier (RM) of the MAXENT model and the importance of add indicators to evaluate model performance, we used ENMeval to improve the MAXENT niche model and conducted an in-depth study on the potential distributions of these two alpine medicinal plants. We adjusted the parameters of FC and RM in the MAXENT model, evaluated the adjusted MAXENT model using six indicators, determined the most important ecogeographical factors (EGFs) that affect the potential distributions of these plants, and compared their current potential distributions between the adjusted model and the default model. The adjusted model performed better; thus, we used the improved MAXENT model to predict their future potential distributions. The model predicted that P. longiflora Rudolph and its variant (P. longiflora var. tubiformis (Klotzsch) Tsoong) would move northward and showed a decrease in extent under future climate scenarios. This result is important to predict their potential distribution regions under changing climate scenarios to develop effective long-term resource conservation and management plans for these species.