Assessment on Potential Suitable Habitats of the Grasshopper Oedaleus decorus asiaticus in North China based on MaxEnt Modeling and Remote Sensing Data

Abstract

Grasshopper populations can quickly grow to catastrophic levels, causing a huge amount of damage in a short time. Oedaleus decorus asiaticus (Bey-Bienko) (O. d. asiaticus) is the most serious species in Xilingol League of the Inner Mongolia Autonomous Region. The region is not only an important grassland but also a site of agricultural heritage systems in China. Therefore, projecting the potential geographic distribution of O. d. asiaticus to provide an early warning is vital. Here, we combined temperature, precipitation, soil, vegetation, and topography with remote sensing data to screen the predictors that best characterize the current geographical distribution of O. d. asiaticus. A MaxEnt model approach was applied to project the potential suitable distribution of O. d. asiaticus in Xilingol League (the Inner Mongolia Autonomous Region of China) combined with a set of optimized parameters. The modeling results indicated that there were six main habitat factors that determined the suitable distribution of O. d. asiaticus such as the soil type (ST), grassland type (GT), elevation, precipitation during the growing period (GP), precipitation during the spawning period (SP), and normalized difference vegetation index during the overwintering period (ONDVI). The simulated result was good, with average AUC and TSS values of 0.875 and 0.812, respectively. The potential inhabitable areas of grasshoppers were 198,527 km², distributed mainly in West Urumqi, Xilinhot City, East Urumqi, Abaga Banner, and Xianghuang Banner of Xilingol League. This study is valuable to guide managers and decision-makers to prevent and control the occurrence of O. d. asiaticus early on and this study may facilitate meaningful reductions in pesticide application.

Keywords: MaxEnt modeling; agricultural heritage systems; grasshopper; habitat factor; remote sensing; suitable area.

Figure 1

The environmental factors and reproductive periods of O. d. asiaticus.

Figure 2

Experimental area location and field survey sampling positions of O. d. asiaticus.

Figure 3

The correlation of each habitat factor (a) and ROC curves in MaxEnt model (b). The red and blue lines are the fit of MaxEnt model for training and testing data, respectively.

Figure 4

Modelling suitability of O. d. asiaticus occurrence by MaxEnt models. The suitability is classified into four levels: high suitability (>0.66, reddish-brown color), moderately high suitability (0.33–0.66, dark yellow color), moderately low suitability (0.05–0.33, grey color), and low suitability (<0.05, blue color).

Figure 5

Relative percent contribution and permutation importance of habitat factors.

Figure 6

Response curves of dominant habitat factors in the model. Response curves show the relationships between existence probability of pest and the six habitat factors. For each panel, the X and Y axes represent habitat factors and probability presence. Red line represents the average of 10 replicates, and the blue margins represent standard deviation of the 10 replicates.