Do Migratory Wildebeest Grazing Influence Soil Organic Carbon Stocks in the Mara–Serengeti Ecosystem?

Each year, over a million wildebeests along with other wild herbivores such as zebras and gazelles move across the Mara–Serengeti ecosystem in a continuous cyclical search for fresh pasture and water, forming one of the most iconic wildlife migrations on Earth. But beyond this spectacular event, something far more critical may be happening beneath their hooves: a process that links directly to the global carbon cycle and by extension the fight against climate change.

Background

Grasslands such as the Mara-Serengeti cover about 40% of the Earth's land surface (excluding Greenland and Antarctica). Soils in these ecosystems play a crucial role in the global carbon cycle and climate regulation by storing soil organic carbon (SOC), the largest terrestrial carbon reservoir. However, in these ecosystems, the loss or accumulation of soil carbon is strongly influenced by vegetation productivity and disturbance regimes, such as grazing, amongst other factors.

The question: Does this cyclical migratory grazing, as a disturbance regime, influence soil organic carbon stocks?

To understand this relationship, it is essential to consider how wildebeest grazing influences vegetation productivity. Through their cyclical grazing, wildebeests remove aboveground biomass, which stimulate new growth, maintaining grasses in an active, productive state. At the same time, nutrient inputs from their dung and urine improve soil fertility, further supporting vegetation productivity. Since SOC is largely derived from plant inputs which include roots, litter, and organic matter, any process that influences vegetation productivity ultimately affect the accrual or loss of SOC in the soil. In this way, wildebeest grazing in the Mara–Serengeti has the potential to indirectly regulate SOC through its effects on vegetation productivity.

Explore the Interactive Dashboard

Toggle layers: grazing intensity · SOC distribution · rainfall · SOC residuals · elevation · Wildebeest GPS sightings across 5 grazing seasons

Open Interactive Dashboard →

Key Findings

SOC in Grazed Areas vs SOC in Ungrazed Areas
51.3 VS 52.4
Mean tonnes of C/ha
Residual SOC in Grazed Areas vs in Ungrazed Areas
+2.16 vs −3.14
Mean tonnes of C/ha

Check out the animated map of the wildebeest migration

Mapped from Collared Wildebeests

Check out the animated map →

Forage Availability and Wildebeest Migration

Rainfall is a dominant factor structuring vegetation growth across the Mara–Serengeti ecosystem. Seasonal patterns in precipitation creates a dynamic availability of forage.

Seasonal Rainfall (mm) vs Mean NDVI · 2019
Peak NDVI follows peak rainfall, reflecting a lagged vegetation response.

In response, wildebeests' movement appear to be closely aligned with the dynamic availability of forage. Wildebeests, which preferentially grazed on short, nutrient-rich grass, appear to track areas that recently received rainfall and concentrate where grass regrowth is highest.

NDVI by Grazing Intensity per Season

"Wildebeests exhibit forage-tracking behaviour, concentrating grazing pressure in areas with most vegetation."

Moreover, this grazing appears to influence how vegetation respond to subsequent rainfall. Areas that experienced prior grazing showed enhanced regrowth once rain fell, compared to areas that were lightly or mildly grazed, or the areas avoided altogether. This suggests that wildebeest grazing does not suppress vegetation growth, but appears to stimulate vegetation productivity by maintaining grasses in an active growth state, indicating a positive feedback between grazing and vegetation dynamics.

Vegetation Regrowth After Rainfall — Dry Season
Areas grazed during the Aug–Oct dry season were revisited in Nov–Dec to assess whether prior grazing stimulated vegetation regrowth following the short rains.
Vegetation Regrowth After Rainfall — Wet Season
Areas grazed during Jan–Mar (peak wet season) were assessed in the following Apr–May (late wet season) to determine whether grazing stimulated or suppressed vegetation growth.

A comparison between the protected area and the surrounding unprotected area further highlighted differences in vegetation productivity. The protected area, which experienced concentrated and consistent grazing by wildebeests, showed a stronger vegetation response to rainfall. In contrast, the unprotected areas exhibited a relatively stable vegetation productivity showing a weaker response to rainfall. This pattern may be associated with land uses such as agriculture or grazing by livestock within the unprotected area around the Mara–Serengeti ecosystem that remain same through the year.

Vegetation Analysis: Protected Area vs Area Beyond Protected Area
The protected area showed a greater seasonal NDVI change (Δ+0.066 units), about 4.5 times more than the surrounding unprotected area (Δ+0.015), suggesting that migratory wildebeest grazing amplifies vegetation responsiveness to rainfall within the protected ecosystem.

So does Grazing Influence Soil Organic Carbon?

At a glance, SOC between grazed and ungrazed areas within the protected area showed no substantial difference.

Mean SOC Stock by Area (t C/ha)
Grazed areas (51.31 t C/ha) and ungrazed areas (52.42 t C/ha) within the protected zone show comparable raw SOC values.

However, elevation emerged as a key confounding factor influencing SOC distribution. Therefore, SOC values were normalised for the effects of elevation, in order to isolate the influence of grazing on SOC stocks within the Mara–Serengeti ecosystem. Grazed areas showed higher residual SOC than ungrazed areas.

Mean Residual SOC in Grazed areas vs in Ungrazed areas (t C/ha)
After controlling for the confounding effect of elevation via OLS regression, grazed areas showed a mean residual SOC of +2.16 t C/ha, while ungrazed areas averaged −3.14 t C/ha.

Grazed areas, which demonstrated enhanced vegetation regrowth, provided a continuous input of organic matter to the soil through both above and below ground biomass. In contrast, areas that were lightly grazed or were never grazed at all showed weaker vegetation responses and therefore, reduced organic matter inputs, which may limit SOC accumulation over time.

A multivariate regression analysis to assess the relative influence of elevation, grazing, and climate (precipitation) on SOC indicated that elevation was the dominant factor (β=0.67).This is because changes in elevation creates stratification of an ecosystem along elevation gradients, creating microclimates which in turn shape local soil temperature, moisture and vegetation composition and as a result causes variabilityin SOC. Grazing emerged as the second strongest factor (β=0.32), demonstrating a significant influence on SOC. In contrast, precipitation showed only a marginal effect (β=0.05), indicating that while it strongly influenced vegetation productivity, its direct contribution to SOC was comparatively limited.

Standardised Effect Size on SOC
Elevation dominates (β=0.67), followed by grazing (β=0.32), which remains statistically significant after controlling for all other predictors. Rainfall contributes only marginally (β=0.05). Overall model fit: R²=0.45.

Limitations

While this study provided insights into how grazing may be influencing SOC in the Mara-Serengeti, several limitations should be acknowledged.

  1. This study focused primarily on wildebeests. However, in the Mara-Serengeti ecosystem, wildebeest migration occurrs in conjunction with other large herbivores', including zebras, gazelles, and buffaloes. These animals differ in grazing behaviour, diet selection, and spatial use of the landscape, and may also contribute differently to vegetation dynamics which affects SOC storage.
  2. The analysis relied on remotely sensed data and modelled global soil carbon dataset, which, although suitable for large-scale assessments, may not fully capture fine-scale spatial variability in SOC. Field-based measurements would have provided higher accuracy and help validate the findings.
  3. Grazing intensity was inferred from wildebeest movement data rather than directly measured in terms of biomass removal. As such, it represented a proxy rather than a direct quantification of grazing impact.
  4. Additionally, while elevation was accounted for as a key confounding variable, other environmental factors such as soil properties, microclimate variations, and land use practices that may also influence SOC were not explicitly accounted for.