Linking land use and the nutritional ecology of herbivores: A case study with the Senegalese locust

Access to high‐quality food is a main driver of population dynamics. For herbivores, protein and carbohydrates are key nutrients that are notoriously variable in plants and are affected by land use. However, few studies have linked foraging decisions and performance in the laboratory to the nutritional landscape available in the field. Oedaleus senegalensis is a non‐model locust, a grass‐feeder, and the main pest of millet, a subsistence crop in the Sahel. In this study, we examined dietary preference and locust performance across a range of protein:carbohydrate ratios using the Geometric Framework methodology. We then applied a fitness landscape approach to visualize these results with the plant nutrient contents available across four land‐use types: millet, groundnut, fallow, and grazed fields. Finally, we contrasted our results with locust distribution in the field. Several locust species (O. senegalensis included) exhibit density‐dependent colour polymorphism; thus, we also reported individual coloration (brown or green). We found that O. senegalensis preferred moderately carbohydrate‐biased food 1:1.6 protein:carbohydrate ratio. All traits recorded (mass gain, development time, growth rate, moult success and performance index) were best near that ratio and declined on either side presenting a ‘hump‐shape’. Fallow fields contained more plants, particularly grasses, that were both abundant and closer to the optimal protein:carbohydrate ratio recorded from the laboratory experiments. When we surveyed O. senegalensis abundance and proportion, we found that they were more numerous in the fallow fields. Brown morph individuals, the ones associated with high density, were proportionally more abundant in fallow fields than green individuals. Our study provides evidence that variation in nutritional landscapes – relative to an herbivore's optimal nutrient balance – is a key driver of herbivore population distribution and abundance, and can be used to predict bottom‐up effects on herbivore species.