Climate‐driven geographic distribution of the desert locust during recession periods: Subspecies’ niche differentiation and relative risks under scenarios of climate change

Abstract

The desert locust is an agricultural pest that is able to switch from a harmless solitarious stage, during recession periods, to swarms of gregarious individuals that disperse long distances and affect areas from western Africa to India during outbreak periods. Large outbreaks have been recorded through centuries, and the Food and Agriculture Organization keeps a long-term, large-scale monitoring survey database in the area. However, there is also a much less known subspecies that occupies a limited area in Southern Africa. We used large-scale climatic and occurrence data of the solitarious phase of each subspecies during recession periods to understand whether both subspecies climatic niches differ from each other, what is the current potential geographical distribution of each subspecies, and how climate change is likely to shift their potential distribution with respect to current conditions. We evaluated whether subspecies are significantly specialized along available climate gradients by using null models of background climatic differences within and between southern and northern ranges and applying niche similarity and niche equivalency tests. The results point to climatic niche conservatism between the two clades. We complemented this analysis with species distribution modeling to characterize current solitarious distributions and forecast potential recession range shifts under two extreme climate change scenarios at the 2050 and 2090 time horizon. Projections suggest that, at a global scale, the northern clade could contract its solitarious recession range, while the southern clade is likely to expand its recession range. However, local expansions were also predicted in the northern clade, in particular in southern and northern margins of the current geographical distribution. In conclusion, monitoring and management practices should remain in place in northern Africa, while in Southern Africa the potential for the subspecies to pose a threat in the future should be investigated more closely.

Incorporating local adaptation into forecasts of species’ distribution and abundance under climate change

“…Very few of the SDM studies we reviewed looked for evidence of adaptation to local climate in the different intraspecific groups. However, a recent trend in SDM papers is to also use occurrence and climate data to test for similarity in the climatic niches of the groups (e.g., Hu et al, ; Ikeda et al, , Meynard et al, ). An important caveat is that patterns of occurrence by themselves do not provide very strong tests of the hypothesis that intraspecific groups differ in their environmental tolerances.…”

“…Other advantages of SDMs are that they commonly consider an extensive set of potential climate variables (Supporting Information Appendix S2), increasingly incorporate model uncertainty by using several algorithms, and build uncertainty in future climate into ensemble-based distribution forecasts (Table 1; Araújo & New, 2007;Thuiller, 2004 Very few of the SDM studies we reviewed looked for evidence of adaptation to local climate in the different intraspecific groups. However, a recent trend in SDM papers is to also use occurrence and climate data to test for similarity in the climatic niches of the groups (e.g., Hu et al, 2017;Ikeda et al, 2017, Meynard et al, 2017. An important caveat is that patterns of occurrence by themselves do not provide very strong tests of the hypothesis that intraspecific groups differ in their environmental tolerances.…”

A young age of subspecific divergence in the desert locustSchistocerca gregaria, inferred by ABC Random Forest

“…In a previous study (Meynard et al 2017), we used Climond data (Kriticos et al 2012) to characterize the current distribution of the two desert locust subspecies, and project their potential fate under different climate change scenarios into 2070. Here, we could not use the same models because the climatic variables used to calibrate those models, in particular radiation and wetness indices, are not available into the mid-Holocene (HCO) and Last Glacial Maximum (LGM) periods.…”

“…For hindcasting, we first need to calibrate a model under current conditions, which can then be used to draw maps of potential ranges under HCO and LGM conditions. To do so, we used the occurrence data published in Meynard et al (2017), which included decades of monitoring of S. g. gregaria in northern Africa, the middle east and southern Asia during remission periods, as well as field records and literature records of S. g. flaviventris in southern Africa. However, rather than modelling each clade independently as we did before, here we decided to group presence records for the two clades as a single unit for the following reasons: (1) here the emphasis is in paleo-climates, at a time scale where the two clades were not yet supposed to be differentiated; (2) a niche analysis in Meynard et al (2017) did not show evidence of niche differentiation between S. g. gregaria and S. g. flaviventris, suggesting that the environment occupied by the southern clade is a subset of that occupied by the northern one; and (3) the biogeographic hypothesis supported by the molecular analysis presented here involves the colonization of the southern tip of south Africa by an isolated population of the northern sub-species, which is more generalist in terms of its environmental niche.…”

“…Finally, since here the emphasis is on Africa, we included only occurrence records in this continent, and used 10,000 pseudo-absences drawn at random from outside the combined current ranges of the two clades by using the distribution maps proposed in Meynard et al (2017) to mask out areas available for pseudo-absence selection. We also assumed here that most of the African continent was accessible to the species during this geological time-scale, given the long known dispersal distances during outbreak periods, the wide distribution of the northern clade, and the fact that the species is present in both extremes of the continent.…”