Desertification Reversal Through Planned Grazing And Earthworks

The United Nations Convention to Combat Desertification (UNCCD) estimates that approximately 12 million hectares of productive land are lost to desertification and drought each year. At that rate, an area equivalent to the size of Greece is lost every twelve months. Over the past 25 years, cumulative losses represent land that could have fed hundreds of millions of people.

The economic cost is quantified at approximately $490 billion annually in lost ecosystem services and agricultural productivity. The social cost — forced migration, conflict over remaining productive land, nutritional collapse — is not easily monetized but is geopolitically visible. The Sahel, the Horn of Africa, the Middle East's Fertile Crescent, and the Loess Plateau in China are all regions where land degradation has directly contributed to political instability.

The global dryland area covers 41 percent of Earth's land surface and is home to approximately 2.5 billion people, many of them among the world's most economically vulnerable. These are not peripheral populations. They are the majority of farmers and pastoralists in Africa, the Middle East, Central Asia, and large parts of South Asia.

To understand why conventional rest-based approaches fail in many dryland grasslands, it is necessary to understand how those ecosystems evolved. The world's major grasslands — the African savanna, the North American Great Plains, the Eurasian steppes, the Patagonian pampas — co-evolved with enormous herds of megafauna. In Africa, this included wildebeest, zebra, buffalo, and their predators. In North America, bison numbering perhaps 30 to 60 million animals ranged continuously, moving in response to predator pressure and forage availability.

These animals had several effects on the grassland ecosystem. Their hooves broke soil surface capping, allowing rainfall to penetrate rather than run off. Their dung transferred nutrients from one part of the landscape to another and provided habitat for insects and fungi critical to decomposition and nutrient cycling. Their grazing removed old and dead plant material, stimulating new growth. And critically, they moved — preventing any single area from being continuously grazed beyond its recovery threshold.

The removal of these megafauna, combined with continuous sedentary grazing by domesticated livestock, broke this cycle. Continuous grazing concentrates animal pressure on the most palatable and accessible areas, preventing recovery. The most productive plants — perennial grasses with deep root systems — are selectively grazed out, replaced by annuals and eventually by bare ground. As root systems die and soil carbon declines, the water retention capacity of the soil falls, reducing moisture availability even in normal rainfall years. A positive feedback loop of degradation is established.

The insight of planned grazing is that the megafauna themselves were not the problem. The removal of natural movement patterns was the problem. Reintroducing movement — whether by wild animals, which is often impossible, or by managed livestock — can reintroduce the biological processes that maintained grassland health.

Allan Savory's work is cited here not uncritically. His TED talk claims — that holistic planned grazing could reverse all of the world's deserts — were received with significant scientific pushback, and some of the early trial data he cited was later questioned. A 2014 review in Rangelands Journal found mixed results across peer-reviewed studies, with some showing vegetation improvement and others showing no significant difference from well-managed conventional grazing.

What the evidence does support, more carefully stated, is this: in seasonally dry grasslands where annual rainfall is 300 to 700 millimeters, planned grazing that provides adequate recovery periods between grazing events can significantly improve vegetation cover, soil carbon, and water infiltration compared to continuous grazing. The effect is not universal and depends heavily on proper implementation, which requires consistent monitoring and adjustment of grazing plans based on actual recovery rates.

The project most often cited as definitive is the work at the Africa Centre for Holistic Management in Zimbabwe, which has documented measurable recovery of vegetation and soil across approximately 65,000 acres. Independent ecologists who have visited the site report real but context-specific results. The evidence from Patagonia, where planned grazing has been implemented across several ranches, and from specific U.S. operations documented by researchers at New Mexico State University, adds to a body of work that is less dramatic than Savory's public claims but more substantive than his critics acknowledge.

The practical conclusion: planned grazing is a tool that works in specific ecological contexts, requires skilled management, and produces results measurable in years to decades rather than seasons. It is not a silver bullet. It is a real and underutilized component of dryland restoration.

Earthworks for dryland restoration operate on a different mechanism than grazing management. Where planned grazing addresses the biological and chemical processes of soil health, earthworks address the physical hydrology — specifically, the problem that degraded soil surfaces repel water rather than absorbing it.

When soil crust forms on bare, compacted ground, the effective infiltration rate can drop to near zero. A 25-millimeter rainfall event on crusted soil generates nearly 100 percent runoff. On well-structured soil with living roots and surface litter, the same rain event may generate 10 to 30 percent runoff, with the rest infiltrating to recharge soil moisture and groundwater. The difference is the difference between a productive agricultural year and a failed one.

Earthworks that slow runoff and concentrate rainfall include:

Contour bunds: Low earthen banks built on the contour of slopes, creating a series of terraced zones that trap runoff and allow it to infiltrate. Used extensively in South Asia and Africa.

Half-moon catchments: Semicircular earthen ridges that concentrate rainfall around a central planting point. Used in the Sahel for both pasture restoration and crop production, with documented yield improvements of 40 to 100 percent in marginal rainfall years.

Rock lines on contour: Simple alignments of stones on slopes that intercept runoff. Extremely low-cost and effective, requiring no machinery. Documented in Burkina Faso to restore vegetation to severely degraded land within three to five years.

Zaï pits: Hand-dug planting holes, typically 20 to 30 centimeters in diameter and 10 to 15 centimeters deep, filled with compost or manure before planting. They concentrate both water and nutrients. The technique originated in Burkina Faso and has spread across the Sahel. It has allowed farmers to bring completely bare, hard-pan soil into crop production.

The Loess Plateau restoration in China is the largest earthworks-based restoration project ever completed. Between 1995 and 2009, the Chinese government, in partnership with the World Bank, invested in a landscape-scale earthworks and revegetation program across approximately 35,000 square kilometers of severely degraded land. Contour terracing, check dams, and controlled grazing combined to reduce erosion from approximately 15,000 tonnes per square kilometer per year to 3,000. Vegetation cover increased from roughly 17 to 34 percent within a decade. River sediment loads declined dramatically. Agricultural productivity on the restored terraces exceeded pre-degradation levels.

Degraded grassland soils have lost between 40 and 70 percent of their original carbon stocks, according to estimates from the IPCC and independent soil carbon researchers. Globally, the world's grasslands and rangelands represent the largest terrestrial carbon store after oceans — approximately 343 gigatonnes of carbon in soil, more than all the carbon in tropical forests above ground.

Even a partial restoration of this lost carbon — recovering half of the depleted fraction across currently degraded drylands — would represent a sequestration magnitude that climate scientists consistently place among the highest-impact land-based mitigation strategies available.

The mechanism is straightforward: living plant roots exude carbon into the soil, feeding microbial communities that stabilize organic carbon compounds into increasingly stable forms. Perennial grasses, which have root biomass roughly equal to or exceeding their above-ground biomass, are among the most effective carbon accumulators in the world when they are allowed to grow and recover between grazing events. Bare or degraded grassland accumulates essentially no carbon. Recovering grassland can sequester 0.5 to 1.5 tonnes of carbon per hectare per year.

At 900 million hectares of degraded dryland with restoration potential, the numbers are in the range of 0.5 to 1.4 gigatonnes of carbon sequestered annually in a recovery scenario — comparable to the total carbon impact of the global aviation industry.

The civilizational plan for dryland restoration requires simultaneous action on four fronts:

1. Tenure and rights: Pastoralists and dryland farmers must have secure rights over the land they manage. Where land tenure is insecure, there is no incentive for long-term investment in soil health. Many of the world's most degraded drylands are communal grazing lands where tenure insecurity creates a tragedy-of-the-commons dynamic. Secure tenure — whether individual, communal, or through formally governed commons — is prerequisite.

2. Knowledge transfer: Planned grazing requires skills that are not self-evident. Monitoring recovery rates, adjusting stocking densities, designing grazing plans around actual rainfall rather than calendar time — these require training and ongoing learning. Investment in extension services that understand and teach these methods is several orders of magnitude less than the cost of dam infrastructure and consistently neglected.

3. Earthwork financing: Simple earthworks can be built with hand tools and local materials but require labor and some organizational coordination. Programs that pay communities for restoration labor — conservation work programs — have been effective in Ethiopia, where the Productive Safety Net Programme employed millions of rural poor to build terraces and water harvesting structures.

4. Market links for restored land: Livestock and crops from restored drylands must reach markets at prices that make continued stewardship economically viable. Carbon markets, ecosystem service payments, and premium pricing for regeneratively produced food are all mechanisms that can provide the economic signal required for long-term investment in soil health.

The degraded drylands of the world are not a fixed feature of the landscape. They are a managed outcome of specific decisions about land tenure, grazing practice, and earthwork investment. Different decisions produce different outcomes. The evidence that this is reversible is strong. The question is whether the will exists to design and fund the reversal at a scale commensurate with the problem.