Niche change analysis as a tool to inform management of two invasive species in Eastern Africa

Abstract

Significant progress has been made in providing guidelines and recommendations for assessing the ecological niche, stage of invasion, and probability of invasive alien plant species (IAPS) potential distribution in space and time. We followed these recommendations by developing and comparing ordination and species distribution models (SDMs) of two important woody IAPS in Eastern Africa, Prosopis juliflora and Lantana camara, and interpreting the results to inform IAPS management. The two species differ in their invasion history in Eastern Africa; while L. camara was widely introduced there in the 19th century, P. juliflora was only planted at selected locations in the 1970s and 1980s. For the SDMs, machine learning algorithms were used to generate one ensemble model each for P. juliflora and L. camara. For ordination, we used bioclimatic variables, performed a principal component analysis, and compared the native and global niches of the species with the Eastern African niche. Niches varied substantially depending on the percentage of marginal climates excluded from the models. Additional analysis of the local niches surrounding the original P. juliflora plantations showed that they are complementary, which may have led to an overestimation of regional niche filling. While niche expansion was absent or small depending on the percentage of marginal climates excluded, analysis of the stages of invasion suggested that P. juliflora may have started to adapt to novel climatic conditions and that L. camara is approaching a pseudo‐stable equilibrium in Eastern Africa. The SDMs showed that large areas in Eastern Africa that have not yet been invaded by P. juliflora are suitable or will become suitable with climate change. For L. camara, the global SDM predicted a considerably larger suitable area than the Eastern African one, raising uncertainty about the areas to be included in a regional management strategy. Thus, combining ordination and SDMs and integrating a geographic component into ordination is useful in assessing IAPS invasion stages and potential niche shifts, and the results help inform IAPS policy and management. The combined approach can also serve to guide experimental studies addressing divergences between results generated with the different approaches.

Stakeholder priorities determine the impact of an alien tree invasion on ecosystem multifunctionality

“…Data were collected in Baringo County, Kenya, a semi‐arid area located in the Great Rift Valley, as part of the larger Woody Weeds project (e.g. Eckert et al., 2020; Linders et al., 2019; Mbaabu et al., 2019). Baringo has a long history of Prosopis invasion and a wide variety of land uses (Mbaabu et al., 2019) and stakeholder groups, making this area especially suited for our study.…”

“…Data were collected in Baringo County, Kenya, a semi-arid area located in the Great Rift Valley, as part of the larger Woody Weeds project (e.g. Eckert et al, 2020;Linders et al, 2019;Mbaabu et al, 2019).…”

Impacts of climate change on current and future invasion of Prosopis juliflora in Ethiopia: environmental and socio-economic implications

“…Again, 500 pseudo-absence points were generated by means of random sampling. To avoid the influence of false absences, we checked and removed points that were closer than 10 km to species presence point following the method of Eckert et al (2020) .…”

Ensemble Models Predict Invasive Bee Habitat Suitability Will Expand under Future Climate Scenarios in Hawai’i

“…Species distribution models (SDM) have been used to compare species' regional ecological niches and forecast the range shifts of species under future climate change scenarios [1,[50][51][52][53]. Species distribution modeling is an approach that predicts the distribution of a species across geographic space and time using the correlation between the geographic occurrence or abundance of a species and corresponding environmental conditions to predict the most suitable habitat [52,54]. Various methods have been used in SDMs, including regression, machine learning, classification, and maximum entropy [48,55].…”