Madagascar’s unique history has created unique restoration challenges

Leighton Reid describes new research linking slow forest recovery to the ancient and protracted isolation that has made Madagascar a hotspot of global endemism – plus an example of working with local farmers to overcome these challenges and restore native rain forest.

Madagascar is a special place with a special history. Separated by ocean from Africa and India for the last 88 million years, this isolated tropical island has fostered the evolution of plants and animals found nowhere else on Earth. Lemurs, couas, and the plant family Sarcolaenaceae are all examples of organisms that evolved only in Madagascar. Collectively, such endemic species make up more than 80% of all plants and animals there.

Crested coua (Coua cristata), one of nine species in the genus Coua – all of which are found only in Madagascar. Photo credit: Olaf Oliveiero Riemer (CC BY-SA 3.0).

Madagascar also has special problems. Almost half of the island’s forest has been cleared for agriculture since 1953, and remaining forests are at imminent risk. One recent study projected that if deforestation rates do not diminish soon, 93% of eastern Malagasy rain forest could be gone by 2070.

The combination of a large proportion of endemic species and a high degree of habitat loss makes Madagascar a biodiversity hotspot. Some people call Madagascar one of the hottest hotspots because its endemism and habitat loss are so extreme.

This week, a new study led by UC Berkeley PhD student Kat Culbertson identified another special problem in Madagascar: following disturbance, Malagasy forests recovery very slowly. Compared to other tropical forests around the world, Malagasy rain forests recover only about a quarter (26%) as much biomass in their first 20 years of recovery. Dry forests in Madagascar also recover more slowly, recovering just 35% as much biomass as American tropical dry forests over the same time period.

Slow biomass recovery following disturbance in Madagascar (dark blue) compared to Central and South America (Neotropics), Africa (Afrotropics), and Asia (Asiatic tropics). Source: Katherine Culbertson et al. (2022) Biotropica.

Why do Malagasy forests recover more slowly than forests in other regions? The answer may be related to Madagascar’s unusual evolutionary history. Culbertson and her co-authors developed four hypotheses and reviewed an array of scientific literature to evaluate support for each one.

Four ways that Madagascar’s unique history could lead to slow forest recovery

1. Native Malagasy forests lack resilience to shifting nutrient and fire regimes from current farming practices. Many rural people across Madagascar practice tavy, a farming method that involves clearing forest, burning it, and then growing rice – a staple crop. After one or a few years of growing rice, the land is allowed to recuperate for several years before it is cultivated again. In other tropical forest locations, such as southern Mexico where humans have farmed for thousands of years, similar practices can coexist with native forests, but Malagasy forests seem to have little resilience to tavy, as least at the intensity with which it is practiced today. For example, in eastern Madagascar, a 3-5 year tavy cycle can cause a native forest to transition to permanent herbaceous vegetation in just 20-40 years. The soil nutrient stocks in that fallow field may be as little as 1-6.5% of soil nutrients stocks in intact forest.

2. Madagascar is an island, and islands tend to have more problems with invasive species. Goats in the Galapagos, brown tree snakes in Guam, acacia in Hawaii, and rats everywhere – these are just some of the ways that island ecosystems have been overwhelmed and transformed by invasive species. Madagascar is no exception. Rain forest regeneration at Ranomafana is stalled by invasive guava, eucalyptus, and rose apple, while dry forest regeneration at Berenty is inhibited by a vine – Cissus quadrangularis. People in Madagascar have many more anecdotes about problems with invasive species like silver oak and Melaleuca quiquenervia, although the extent and impact of these invaders on forest recovery have not yet been studied.

3. Old, weathered soils have favored the evolution of slow-growing native plants. Madagascar is not only an island, it is a very old island, and as such its soils have been weathered and depleted of important nutrients like phosphorus. It’s hard to separate the effect of inherently low nutrient availability due to being an old island from the effect of human-induced nutrient scarcity through tavy, but one comparison of phosphorus content in rice stalks showed that phosphorus content was 10× lower in Madagascar compared to the rest of sub-Saharan Africa. If native trees have evolved to grow more slowly in Madagascar because of low nutrient availability, then on average exotic tree species should grow faster than native Malagasy ones in the same gardens. This has been shown in a few cases, but a more compelling analysis would need more species.

4. Finally, Malagasy forests have dysfunctional seed dispersal. One way in which Madagascar is different from other tropical areas is that by and large its trees have evolved to have their fruits dispersed by lemurs. Unfortunately, many of the lemurs that could disperse Malagasy tree fruits are either extinct or endangered – in many cases due to a combination of hunting and habitat loss. Moreover, the lemurs that remain are reluctant to venture outside of forest fragments (perhaps with good reason) and so they are unable to disperse seeds to regenerating farmlands that most need them.

Black and white ruffed lemur (Varecia variegata) – a critically endangered seed disperser in eastern Madagascar. Photo credit: Tim Treuer.

In essence, the ancient and protracted isolation that has made Madagascar so unique has also made it uniquely vulnerable to contemporary changes like deforestation, fire, and agriculture. The result is an unfortunate combination: Madagascar not only has some of the highest deforestation rates, it is also one of the places least ecologically equipped to rebound from those disturbances.

A mosaic of mature tropical dry forest and forest restoration at Berenty in southern Madagascar. Photo credit: Ariadna Mondragon Botero.

The way forward – working with local people

Despite these challenges, Madagascar has committed to restoring four million hectares of lost habitat by 2030, an area nearly 7% the total national territory. This is a tall order in a country where technical difficulties are high and financial resources are often low, but it can be done, and the way forward, undoubtedly, is to work with local people.

One group that exemplifies bottom-up restoration is GreenAgain, a non-profit restoring native rain forest and supporting rural livelihoods in eastern Madagascar. GreenAgain is led and staffed by farmer-practitioners whose neighbors, family, and friends contract with GreenAgain to design, plant, and monitor diverse native forests on their lands. Last year, GreenAgain staff planted 20,000 trees across central eastern Madagascar, each one carried by hand, on foot, from one of eight regional tree nurseries. The rural farmers at GreenAgain collect rigorous data on tree survival and growth and collaborate with scientists to analyze and share the results of their tree planting experiments.

For example, one of the earliest experiments at GreenAgain was an assay of tree planting strategies intended to improve native tree seedling survival during plantings that occur in the dry season. Trees planted during the dry season typically have high mortality, sometimes in excess of 40%. One of the strategies that local farmers recommended to improve survival was to erect small teepees over each seedling using the leaves of a common fern, Dicranopteris linearis. These structures are temporary – they eventually dry out and blow away – but GreenAgain’s experiment showed that they reduced transplant shock (i.e., mortality in the first few weeks) by 75% compared to seedlings that were left to bake in the hot sun. In contrast, many of the other treatments had no discernable effect.

To analyze and publish these findings, GreenAgain partnered with an award-winning undergraduate researcher, Chris Logan, in my lab at Virginia Tech, who led a peer-reviewed paper that is now available at Restoration Ecology.

Leaf tent made with a ubiquitous fern, Dicranopteris linearis, placed over a native tree seedling. Photo credit: Catherine Hill.

Could technological solutions like hydrogels or irrigation systems produce greater improvements in dry season tree survival? Yes – they probably could for a certain price, but homegrown solutions like fern leaf shade tents are free and easily accessible to any person doing restoration across eastern Madagascar. They are also more likely to be used because they were developed by local people.

This study also showed that some native tree species are much better at coping with dry season stress than other species, so another possible solution for dry season plantings could be to plant only the tough survivors. Once those trees survive and begin to produce shade, fern leaf tents may not even be needed anymore to help more sensitive native species survive and grow.

To read more about ongoing restoration and ecological research in Madagascar, read our new review of how Madagascar’s evolutionary history limits forest recovery and our new open-access paper about strategies for dry season plantings in eastern Madagascar.

If you are in a position to support the work of local farmers restoring rain forests in eastern Madagascar, consider donating to GreenAgain at their website,

Global pledges to restore forests face challenges, and need increased support

Matthew Fagan is an assistant professor in Geography and Environmental Systems at University of Maryland Baltimore County. Here he describes the challenges confronting countries as they attempt large-scale forest restoration, and why many countries will need help to fulfill their goals. For more information, read his new, open-access paper in Conservation Letters.

Degraded and deforested landscapes are widespread, and tropical forests are being lost at a rate of 15.8 million hectares a year. But there is good news—temperate forest area is increasing, and more and more countries are voluntarily pledging to restore vast tracts of degraded land. Restoring forests benefits biodiversity and society, and can combat global warming as well, as growing trees lock away carbon dioxide.

International interest in restoring trees to landscapes emerged out of policy discussions last decade, and resulted in the 2011 Bonn Challenge and the creation of voluntary national restoration targets by many countries. The Bonn Challenge seeks to bring 150 million hectares into restoration by 2020, and 350 million hectarees by 2030 (that’s roughly 700 million American football fields, 350 million rugby fields, 500 million FIFA football fields, or an area a bit larger than India).

Current Bonn Challenge pledges total some 172 million hectares. That’s a massive international commitment, and when you add in internal commitments by countries, the potential restoration area swells to 318 million hectares.

All that area voluntarily committed to restoration got my co-authors and I excited, but also skeptical—were countries really going to follow through on their commitments?


A rain forest blow-down in northeastern Costa Rica, with a storm-downed tree cut to clear a path. Silviculture restoration promotes the recovery of disturbed forests like this one. Photo credit: Matthew Fagan.

To try to answer that question at this early stage, myself, Leighton Reid (Virginia Tech), Maggie Holland (UMBC), Justin Drew (UMBC), and Rakan Zahawi (University of Hawaiʻi at Mānoa) asked three related questions in a recent paper in Conservation Letters.

  1. Is the amount of land a country pledged to restore related to their past record of restoring forested landscapes and implementing sustainable development?
  2. For the small group of countries that have publicly reported their progress on commitments, is the amount of restoration they completed predictable by their development level or other risk factors, like deforestation?
  3. Which countries will likely face the greatest challenges to meet their commitments and maintain restored land into the future?

We then set to gathering published information on country commitments and progress, and recent national rates of forest loss, agricultural expansion, and forest recovery.


Recent natural regeneration in northeastern Costa Rica of varying ages. Photo credit: Matthew Fagan.

All of these programs seek to reforest landscapes in ways that benefit both nature and people, including options like natural regeneration (letting natural forests recover and expand), silviculture (interventions to restore standing forests, like preventing forest fires and promoting recovery from selective logging), tree plantations (often tree monocultures to produce timber and pulp on degraded lands), and agroforestry (planting trees on and around farmland to shade crops or protect streams and fields). These options are not all equal in their benefits for biodiversity, carbon, and society, but a diverse menu of options allows countries to consider committing to at least some form of restoration over large areas.


A tree plantation in northeastern Costa Rica funded by the national payments for environmental services program. It is a monoculture of a single native species, Vochysia guatemalensis, grown for timber. Photo credit: Matthew Fagan.

In a nutshell, what we found was both discouraging and encouraging.

First, after adjusting for the size of a country and how much restoration they had done previously, we found that less-developed countries committed more land for restoration. This might be for positive reasons; for example, they may be taking proactive action against the greater risk they face from climate change. Or it might be because they underestimated how challenging it would be to achieve a large pledge.


Silvopastoral restoration, a type of agroforestry, in northeastern Costa Rica. The understory is a cattle pasture, while the overstory is plantation of a native tree species, Dipteryx panamensis. Photo credit: Matthew Fagan.

Second, for twelve early-reporting countries, restoration progress was predictable based on a risk index. Countries with higher risk (risk factors included deforestation rates and progress on sustainable development goals, among others) had less restoration progress.

Third, countries made massive individual commitments that will be hard to achieve without wholesale transformation of their food systems. One third of countries committed >10% of their land area (with a maximum of 81%, in Rwanda). A quarter either committed more area than they had in agriculture, or committed more area than they had in forest. And one quarter of countries had more forest loss and agricultural conversion in 2000–2015 than their restoration commitment for 2015–2030.


Coffee plantation under tree cover, a type of agroforestry, in central Costa Rica. The understory is a monoculture of coffee shrubs, while the overstory is scattered planted trees. The partial cover helps the shade-loving coffee plants stay healthy, but many coffee farmers are moving away from this traditional farming approach. Photo credit: Matthew Fagan.

As noted in our paper, “If voluntary commitments like the Bonn Challenge fail to precipitate meaningful restoration across large areas, the UN’s vision of a sustainable future will become less attainable.” But what this study found is not countries that have failed on their restoration pledges. We are still in the first days of the UN Decade of Ecosystem Restoration. What we have identified is countries that will need help to restore their lands.

We believe it is time for the international community to step up and aid all countries in achieving their restoration goals. To quote Thoreau, “If you have built castles in the air, your work need not be lost; that is where they should be. Now put the foundations under them.”


A regrowing forest in central Costa Rica, showing the promise of restoration. Photo credit: Matthew Fagan.

The ephemeral forests of southern Costa Rica

Damaged ecosystems don’t recover overnight, but sometimes that’s all the time that they get. CCSD scientist Leighton Reid describes new research about tropical secondary forests in southern Costa Rica, including how long these young forests last, what’s at stake, and how we can keep them around longer.

Regrowing tropical forests on marginal farm lands is one of the main ways that humans can prevent runaway climate change. With ample moisture and long growing seasons, tropical trees often can grow quickly and pull large amounts of carbon out of the atmosphere, storing it in their wood and keeping it from trapping heat. At the same time, young forests provide habitat for plants and animals and improve water quality for humans, among many other benefits.

But even in a moist, tropical climate, trees don’t grow instantly. Typically, it takes many decades for a recovering forest to stock up all of the carbon that it can hold. And it can take even longer for some plants (like orchids) and animals (like antbirds) to return. If a forest starts to grow back, but then someone cuts it down again, these time-dependent benefits never accrue.

In other words, the hopes and expectations that many people have for young tropical forests depend on young tropical forests growing old. So do they? Our new study suggests not.

San Vito & Coto Brus Valley

The Coto Brus Valley and Talamanca Mountains in southern Costa Rica. Photo by J. Leighton Reid.

To find out how long secondary forests persist, I teamed up with Matthew Fagan, a landscape ecologist at the University of Maryland Baltimore County, and Rakan Zahawi, director of the Lyon Arboretum, as well as two students, James Lucas at Washington University and Joshua Slaughter at UMBC.

We studied a set of historical, aerial photos from southern Costa Rica, which covered the time period from 1947-2014. Previously, Zahawi and colleagues had classified which areas in each photo were forest and which areas were farms or other non-forest land uses. By comparing the maps they made for each year, we were able to see where and when new forests appeared and how long they remained as forest before they were converted to some other land use (mostly farms).

The young forests did not last long. Half of the new forests disappeared before they were 20-years old. And 85% were cut down before they were 54-years old. Larger forests and forests near rivers lasted longer.

One hectare forest fragment, Coto Brus, Costa Rica

An isolated forest fragment surrounded by cattle pastures in southern Costa Rica. Photo by J. Leighton Reid.

First, the bad news. Twenty years is not even close to the amount of time it takes for a young forest to become as diverse as an old-growth forest. For example, vascular epiphytes like orchids and bromeliads take more than 100 years to fully recover in young forests.

Carbon storage will also take a hit. If forests elsewhere in Latin America are as ephemeral as forests in southern Costa Rica, then carbon stocking over the next thirty years may be reduced by an order of magnitude.

Ephemeral forests could just be a problem in Costa Rica, but another study shows that secondary forests in eastern Peru have even shorter lifespans. There, secondary forests are cleared at a rate of 3-23% per year. Compared to that, the 2-3% per year rate of loss in southern Costa Rica is considerably better. And that’s not a good thing. Clearly we need more research on secondary forest persistence from other places.

There is some good news, though. Even though many new forests were short-lived, the ones that survived were predictable. And if we can predict where new forests will survive, we should also be able to help them survive longer. Larger forests and forests close to rivers were cut down less often than small forests and forests far from rivers. This suggests that restoring large, riparian forests could be a smart investment.

Gulfo Dulce from Fila Cruces - Coto Brus, Costa Rica

Forests and cattle pastures in southern Costa Rica. Photo by J. Leighton Reid.

Governments and other organizations can also help forests persist by creating incentives for long-term forest management, providing resources to enable long-term management, and ensuring that local people will be able to enjoy the benefits that old forests provide.

We hope that this work will lead to stronger restoration commitments. Right now, dozens of countries are setting big targets for forest restoration. For example, in 2012 Costa Rica committed to restore a million hectares of degraded land by 2020 (an area about one fifth the size of the country). There is a great opportunity for Costa Rica and other ambitious countries to plan for long-term forest restoration.

If we can begin to restore a million hectares of forest by 2020, why not plan to restore a million hectares of 100-year old forest by 2120?

Melissa's Meadow, Las Cruces Biological Station, Costa Rica

A trail through secondary forest at the Las Cruces Biological Station in southern Costa Rica. Photo by J. Leighton Reid.

For more information on this research, you can read our open-access paper in Conservation Letters or watch a video of Leighton Reid presenting to the Association for Tropical Biology and Conservation back in June. Additional papers on restored ecosystem persistence are available here and here. This work is a product of the PARTNERS (People and Reforestation in the Tropics: a Network for Research, Education, and Synthesis) Working Group on Spatial Prioritization. Funding was provided by grant DEB-1313788 from the U.S. National Science Foundation’s Coupled Human and Natural Systems Program.