Figure 3 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
Accuracy evaluation of eleven single models: (A) natural environmental; (B) natural environmental + human activity.
Source publication
Effective use of species distribution models can assess the risk of spreading forest pests. In this study, based on 434 occurrence records and eight environmental variables, an ensemble model was applied to identify key environmental factors affecting the distribution of Apriona rugicollis Chevrolat, 1852 and predict its potential habitat and its r...
Context in source publication
Context 1
... results show that when using only natural environment factors, ANN, GAM, and SRE were excluded from the ensemble model due to their low TSS and AUC values, and we chose RF, CBM, CTA, XGBOOST, MAXENT, MARS, FDA, and GLM to construct the ensemble model. When adding human activity variables, we chose ANN, RF, CBM, CTA, XGBOOST, MAXENT, MARS, FDA, and GLM to construct the ensemble model (Figure 3 When considering only the natural environment factor, the ensemble model had an AUC value of 0.99 and a TSS value of 0.89. However, after adding the human activity variable, the AUC value was 0.98, and the TSS value was 0.87 (Table 2). ...
Citations
... The principle of the MaxEnt model is to calculate the optimal distribution of a species under the constraints of a certain ecological niche, that is, the possible distribution of the species in a certain prediction area when entropy is maximized, based on the principle of climatic similarity and the corresponding environmental variables of the species (Harte et al. 2021;Soliman et al. 2023). Examples include predicting potential distributions of species of different taxa (e.g., mammals, birds, amphibians, and insects), center-ofmass migration, and determining fitness intervals for environmental variables using response curves (Lee et al. 2021;Meza Mori et al. 2023;Zhang et al. 2024d). Therefore, the MaxEnt model is widely used in the prediction of potential distribution areas of rare and endangered species with little distribution data and narrow distribution ranges . ...
... Higher isothermality and lower temperature seasonality may provide Pseudoechthistatus with more stable temperature environments, which in turn reduces physiological stresses due to drastic temperature fluctuations during development. Whereas in areas with high temperature fluctuations, insects may face higher energy expenditure to adapt to rapidly changing environments, stable temperature conditions may favor their growth and development and population reproductive capacity (Zhang et al. 2024d). Moreover, isothermality and precipitation of driest quarter indirectly affect the population dynamics of host plants by regulating their water status and food availability. ...
Climate change will lead to changes in biological ecosystems, which may affect the geographic distribution of Pseudoechthistatus and thus alter the extent and spatial pattern of its habitat. Pseudoechthistatus plays an important role in biodiversity and has significant ecological value. This study utilized an optimized MaxEnt model to predict the predicted distribution of Pseudoechthistatus in China for the current and future (2050s and 2070s). The results show that the MaxEnt model has high prediction accuracy with AUC values higher than 0.97 for both training and testing. The most influential factors contributing to the distribution of Pseudoechthistatus were temperature seasonality (Bio4) and isothermality (Bio3), accounting for 38.8% and 28.2%, respectively. Furthermore, southern China remains a region of high suitability for Pseudoechthistatus species diversity. However, the Beijing climate center climate system model (BCC‐CSM2‐MR) predicts a decrease in suitable areas for Pseudoechthistatus, while the model for interdisciplinary research on climate (MIROC6) predicts an increase in medium and low suitable areas for Pseudoechthistatus. Additionally, future climate change will significantly alter its distribution pattern, with Pseudoechthistatus predicted to decrease its suitable area by 6.64%–28.01% under the BCC‐CSM2‐MR model and increase its suitable area by 6.14%–18.61% under the MIROC6 model. The results show that the MaxEnt model can improve the understanding of the geographical distribution of Pseudoechthistatus in the context of climate change and provide a scientific basis for the identification of potentially suitable habitats and the development of stable suitable areas for conservation.
