Summary1. Large herbivores play a key role in creating spatial heterogeneity through the formation of grazing lawns. Recent research suggests that the currently accepted nutrient-based theory on the formation of these grazing lawns cannot universally explain their formation in all ecosystems where they are found. 2. We developed and investigated an alternative hypothesis on grazing lawn formation and maintenance based on herbivore effects on the plant-soil water balance. We propose that large herbivores change the soil water balance in grazing lawns through defoliation and soil compaction, causing a shift in vegetation composition towards a drought-tolerant plant community. 3. Investigating this idea in a tropical savanna, we indeed found profound differences in grazing lawn soil properties and water balance. In particular, defoliation increased soil temperatures and potential evaporation rates while soil compaction increased bulk density and decreased water infiltration rates, especially on fine-textured soils. Soil moisture was therefore generally much lower in grazing lawns than in adjacent bunch grass areas. 4. Furthermore, we found that grazing lawn species show drought-tolerant traits, with higher leaf sodium levels, suggesting evolutionary adaptation to these herbivore-induced dry conditions. However, leaf water potentials did not differ between grazing lawn and bunch grass species. 5. Synthesis. This study shows that large herbivores might form grazing lawns through previously underestimated effects on water balance. Thus, future studies on large herbivore effects on vegetation should increasingly focus on additional pathways of soil compaction and defoliation. While nutrient-based processes driving grazing lawn formation may operate during the wet season in savannas, we suggest that water balance-based processes are additionally important during the dry season.

Abstract. Grazing ecosystems ranging from the Arctic tundra to tropical savannas are often characterized by small-scale mosaics of herbivore-preferred and herbivore-avoided patches, promoting plant biodiversity and resilience. The three leading explanations for bistable patchiness in grazed ecosystems are (1) herbivore-driven nutrient cycling, (2) plantgrowth-water-infiltration feedback under aridity, and (3) irreversible local herbivore-induced abiotic stress (topsoil erosion, salinity). However, these insufficiently explain the high temporal patch dynamics and wide-ranging distribution of grazing mosaics across productive habitats. Here we propose a fourth possibility where alternating patches are governed by the interplay of two important biotic processes: bioturbation by soil fauna that locally ameliorates soil conditions, promoting tall plant communities, alternating with biocompaction by large herbivores that locally impairs soil conditions, and promotes lawn communities. We review mechanisms that explain rapid conversions between bioturbation-and biocompaction-dominated patches, and provide a global map where this mechanism is possible. With a simple model we illustrate that this fourth mechanism expands the range of conditions under which grazing mosaics can persist. We conclude that the response of grazing systems to global change, as degradation or catastrophic droughts, will be contingent on the correct identification of the dominant process that drives their vegetation structural heterogeneity.

Question: How multiple abiotic stress factors combined with herbivory affect interactions within plant communities is poorly understood. We ask how large herbivore grazing affects the direction of plant-plant interactions along an environmental gradient in a salt marsh.Location: Grazed (cattle) and ungrazed salt marshes of the Dutch Wadden Sea island Schiermonnikoog. Here, patches of tall plant communities, dominated by the tough, unpalatable species Juncus maritimus Lam., are found alternating with low-statured, intensively grazed plant communities.Methods: Along the inundation gradient, we measured plant species composition and plant species traits (specific leaf area, specific root length, maximum height and abundance) inside and outside J. maritimus patches in grazed and ungrazed areas. In addition, we measured soil structure parameters (bulk density, soil porosity, clay depth), multiple limiting conditions for plant growth (soil salinity, soil redox, plant canopy light interception), plant biomass, presence of herbivores and abundance of soil macro-detritivores.Results: Under grazing, the palatable grasses Elytrigia atherica (Link) Kergu elen and Festuca rubra L. were positively associated with J. maritimus, while shadeintolerant Puccinellia maritima (Huds.) Parl. and Juncus gerardii Loisel. were negatively associated with this species. Furthermore, macro-detritivore presence was higher inside J. maritimus patches. In ungrazed areas E. atherica and F. rubra were negatively associated with J. maritimus, while P. maritima and J. gerardii were rare. In both grazed and ungrazed conditions the directions of species associations were independent of the inundation gradient. Analysis of species traits and abiotic conditions suggested that associational resistance (a facilitation type) was important in grazed areas. In ungrazed areas, light competition was the likely dominant process.Conclusions: The direction of species associations within these salt marsh communities was strongly affected by grazing, not by the underlying stress gradient. Measurement of species traits indicated that plant-plant interactions shifted from competitive to facilitative under grazing. Besides grazing, cross-trophic facilitation of soil disturbing macro-detritivores may play an importantthus far ignoredrole in structuring plant communities.

Questions In African savannas, Macrotermes termites contribute to small‐scale heterogeneity by constructing large mounds. Operating as islands of high nutrient and water availability and low fire frequency, these mounds support distinct, diverse communities of trees that have been shown to be highly attractive to browsers. However, the distinct traits of tree species on termite mounds have hardly been studied, even though this may help to understand processes determining (1) their characteristic community structure and (2) attractiveness for browsers. Here, we compare functional trait and browser preference values between tree species on and off termite mounds. Location Hluhluwe‐iMfolozi Park, Kwazulu‐Natal, South Africa. Methods We recorded tree community compositions for 16 large Macrotermes natalensis mounds and 16 control plots of 100 m2 each in a paired design. For each observed tree species we measured 22 traits, related to water and nutrient use, fire tolerance, light competition and anti‐herbivore defence, and compared average trait values between mound and control communities. Furthermore, we investigated the feeding preferences of ungulate browsers for the most common tree species and how this was linked to their associated traits. Results Termite mounds supported tree communities that were distinct from the surrounding savanna vegetation. Mounds hosted more evergreen and less leguminous tree species than control communities, and the dominant species were less mechanically defended, less nutritious, had larger leaves and lower wood density than the species dominating control plots. Browsers preferred leguminous tree species with high leaf N and P content, which were relatively rare on termite mounds. Conclusions Overall, we conclude that termite mounds in this savanna form small refuges for tree species that seem less adapted to fire (more evergreens), have low nutrient availability (less nitrogen fixers) and suffer from water stress (larger leaf sizes) than typical savanna trees. Surprisingly, despite their reputation as browsing hotspots, the tree species dominating mounds are less nutritious and less preferred by browsers than tree species of the surrounding savanna, which may be explained by the relatively nutrient‐rich nature of this savanna or intraspecific trait differences.

Few studies have been able to directly measure the seasonal survival rates of migratory species or determine how variable the timing of migration is within individuals and across populations over multiple years. As such, it remains unclear how likely migration is to affect the population dynamics of migratory species and how capable migrants may be of responding to changing environmental conditions within their lifetimes. To address these questions, we used three types of tracking devices to track individual black-tailed godwits from the nominate subspecies (Limosa limosa limosa) throughout their annual cycles for up to 5 consecutive years. We found that godwits exhibit considerable inter-and intra-individual variation in their migratory behavior across years. We also found that godwits had generally high survival rates during migration, although survival was reduced during northward flights across the Sahara Desert. These patterns differ from those observed in most other migratory species, suggesting that migration may only be truly dangerous when crossing geographic barriers that lack emergency stopover sites and that the levels of phenotypic flexibility exhibited by some populations may enable them to rapidly respond to changing environmental conditions.