Global spatial distribution of Chromolaena odorata habitat under climate change: random forest modeling of one of the 100 worst invasive alien species

Abstract

Anthropogenic activities and global climate change increase the risk of Chromolaena odorata invasion and habitat expansion. To predict its global distribution and habitat suitability under climate change, a random forest (RF) model was employed. The RF model, utilizing default parameters, analyzed species presence data and background information. The model revealed that the current spatial distribution of C. odorata covers 7,892,447 km². Predictions for 2061- 2080 indicate expansion of suitable habitat (42.59 and 46.30%), reduction of suitable habit (12.92 and 12.20%), and preservation of suitable habitat (87.08 and 87.80%) under the SSP (Shared Socio-economic Pathway) 2-4.5 and SSP5-8.5 scenarios, respectively, in comparison to the present distribution. Currently, C. odorata is predominantly found in South America, with limited presence in other continents. However, the data suggest that climate change will elevate the global invasion risk of C. odorata worldwide, particularly in Oceania, Africa, and Australia. Countries such as Gambia, Guinea-Bissau, and Lesotho, which currently have unsuitable habitats, are predicted to have highly suitable habitats with climate change, supporting the idea that global habitat expansion for C. odorata will occur due to climate change. This study indicates that proper management of C. odorata is crucial during the early invasion phase.

Figure 1

Jackknife test showing relative importance of bioclimatic variables used in RF model. The details of variables bio1, bio2, bio3, bio12, bio13, and bio 14 are presented in Table 1.

Figure 2

Global occurrence records of C. odorata (n = 4801). The red points in the map indicate GIS points for C. odorata. Generated using ArcGIS Desktop 10.8 (https://desktop.arcgis.com).

Figure 3

Potential global binary distribution of C. odorata under the current climate (1970–2000) (A) and under future climate change scenarios SSP2-4.5 (B) and SSP5-8.5 (C) (2061–2080). Orange indicates suitable habitat for C. odorata, and white indicates unsuitable habitat. Generated using ArcGIS Desktop 10.8 (https://desktop.arcgis.com).

Figure 4

Changes in habitat suitability for C. odorata under the climate change scenarios SSP2-4.5 (A) and SSP5-8.5 (B) by 2061–2080. Blue indicates habitat expansion, green indicates habitat loss, orange indicates unchanged habitat, and grey indicates unsuitable habitat. Generated using ArcGIS Desktop 10.8 (https://desktop.arcgis.com).

Figure 5

Mean habitat suitability of C. odorata estimated for different countries of the world. (A) Mean habitat suitability under the current climate. (B and C) Mean future habitat suitability under climate change scenarios SSP2-4.5 (B) and SSP5-8.5 (C) by 2061–2080. Gray indicates unsuitable habitat, green indicates habitat with low suitability, yellow indicates moderate suitability, and red indicates high suitability. Generated using ArcGIS Desktop 10.8 (https://desktop.arcgis.com).

Figure 6

Estimation of habitat suitability in different countries of the world under the climate change scenarios SSP2-4.5 (A) and SSP5-8.5 (B). The average habitat suitability of each country was classified as unsuitable (0), low (0.1–0.33), moderate (0.34–0.66), or high (0.67–1.00). Green indicates the number of countries currently in each category; red indicates the predicted number of countries in each category in 2061–2080.