An integrative paper published in Science in 2019 demonstrates that pressures along the edges of the ecosystem, especially by illegal grazing of livestock in protected areas has far-reaching consequences that extend to key ecological processes at the core of the ecosystem.

The land cover change in the GSME over the past 40 years shows rapid changes from pastoral areas to cropland with associated negative consequences for wildlife, especially with higher rainfall.  

Pastures in the livestock-dominated areas outside the park and wildlife-dominated areas inside the protected areas have similar grass productivity.  Grass nutrient quality is higher on pasturelands but is reduced by high rainfall.  or in wetter regions.  Overgrazing is more detrimental to pasture productivity during the dry season.

Below ground nutrient cycling in the form of root decomposition is weakly influenced by land-use but positively influenced by termites and other macro-fauna.  

Carbon storage is similar inside protected areas versus livestock pasturelands.  Almost all total ecosystem carbon (>97%) was found in the soil carbon and this did not differ across the park boundary. Soil carbon however shows signs of being negatively affected by increasing livestock numbers.

There is pronounced rainfall variation, which influences wildlife and livestock losses.  In particular, quasi-periodic oscillations in rainfall have important implications for recurrent severe droughts.

Increased livestock grazing along the margins of the ecosystem causes different degrees of bush encroachment through fire suppression. Bush encroachment depends on the combination of soil fertility, rainfall, and associated woody species adaptations to fire and herbivory.

Severe droughts reduce fecundity and synchrony of reproduction in large mammalian herbivores and delays the onset of their reproduction and timing of birth peaks.  High rainfall increases fecundity and synchrony of reproduction and advances the onset of births in large mammalian herbivores.

The species diversity of small mammals and selected birds and larger mammals was higher inside than outside Serengeti National Park (SNP). Inside SNP, the diversity and abundance of non-migratory herbivores was generally as high or higher than recorded with similar methods 15 years ago.

In areas surrounding SNP, interviews with local residents disclosed that birds are mostly hunted for food, and that some species are hunted regardless of whether they know that hunting protected species is illegal.

AfricanBioServices analysis of wildlife contributed to the IUCN placing the Maasai giraffe as a highly endangered species.

Wildebeest, zebra, impala, and African wild dogs were collared to track migratory movements and home range utilization.  Wild dog distribution was strongly influenced by both lion density as well as human and livestock densities outside the national park.

Wildebeest are more stressed when in human dominated areas, whereas impala stress levels were influenced by forage quality and not necessarily indicative of increasing levels of human disturbance.  

Only the Serengeti-Mara wildebeest migration confined entirely within a protected area is facing relatively lower threat level. All other East African migrations were found to be close to or have reached extinction.  Drivers of the ungulate migration and the home range utilization were forage quality and avoidance of humans.  

Livestock densities have exceeded a threshold to facilitate and coexist with wildlife.  Livestock compete with wildlife, especially with the larger wildlife species.

In the GSME, human reliance on income derived from the environment equals or exceeds global averages. Human well-being was found to increase with distance to the protected area boundary.  

Households overwhelmingly prefer expansion of traditional activities (agriculture, livestock) with increased market access due to road construction.  Hence, road development may further increase ecosystem pressure through land conversion and illegal grazing.

An increase of substitute meats does not reduce the demand for bushmeat. Supply-side interventions such as patrolling and law-enforcement reduce illegal hunting and increase the price of bushmeat, which may influence demand.

A board game simulating potential consequences of various livelihood strategies in the GSME and illustrating the problems of high population growth rate and other drivers of change was developed and played with local stakeholders, policy administrators, and park managers in 12 villages. Women were particularly successful in the game and, even more so than more experienced stakeholders.

Data on human population density, settlements, wildlife and livestock densities, rainfall, primary biomass production, agricultural practices, and socioeconomic data from household surveys have been used to model future land-use, ecosystem integrity and human welfare outcomes contingent on various population, climate and land use scenarios between 2015 and 2030.