Native tree seedlings grow best near existing forest and beneath shade in highland Madagascar

A team of MBG scientists describes a recent experiment to grow native trees in a degraded part of Madagascar’s central highlands.

Tampo_Raw

The dissected landscape of the Tampoketsa de Ankazobe in central Madagascar. Imagery: Google Earth (2018).

Seen from space, parts of Madagascar’s high plateau look like a wizened, grayish-pink brain drying in the sun. Thin, dark lines demarcate nooks and crannies – nearly the only places where bits of forest remain.

Formerly, the forests here covered more territory. Just how much territory is debated; ancient grasslands are also present in highland Madagascar. But in this area, about three hours northwest of the capital, many forests have been cleared, burned, and converted to new grassland within living memory.

To restore forests to their recent extent would benefit a range of species, including Schizolaena tampoketsana (a threatened, micro-endemic tree) and an undescribed species of fat-tailed dwarf lemur. However, restoration has been easier said than done so far. Natural forest regeneration is slow to non-existent, even near remnant forests where fire is excluded. Planted tree seedlings grow only millimeters each year, if they survive at all. Adding fertilizer seems to inhibit seedling growth. Inoculating seedlings with mycorrhizal fungi seems promising, but we are not yet sure if this will make a difference in the field.

Following a field trip in November 2016, we decided to test a couple of other tactics for growing native trees on this weathered plateau. First, we tested planting trees near existing forest. Being near the forest could help young seedlings by shading them from the hot sun or by sharing beneficial microorganisms. Second, we put shade structures over some of tree seedlings to test how much the bright, hot sunlight prevented tree growth.

We tested four tree species: Baronia taratana (Anacardiaceae), Nuxia capitata (Stilbaceae), Uapaca densifolia (Phyllanthaceae), and Eugenia pluricymosa (Myrtaceae).

ExpDesign

An experiment with native tree seedlings at Ankafobe, a small forest fragment on the highlands northwest of Madgascar’s capital, Antananarivo. Photo: Chris Birkinshaw.

Three of the tree species survived and grew more when we planted them next to the forest. The fourth, Nuxia capitata, was a super species and grew relatively well wherever it was planted.

Two of the four species also survived more often beneath shade structures. But interestingly, this shade effect did not completely account for the effect of proximity to forest. That suggests that shade is important for protecting young seedlings from the hot sun, but something else is going on too. Perhaps trees growing next to the forest get a boost of water, since remnant forests sit at the valley bottom where water collects. If this is true, then tree seedlings might do well in any valley bottom, not just ones with remnant forest in them.

Our study site, called Ankafobe, is only a small area, so it would be a stretch to generalize our observations to the entire region. However, we are not the only ones to have done such a test in this ecosystem. In 2000, Ingar Pareliussen led a study with the same basic elements as ours at a site ten kilometers away, at Ambohitantely. Like us, Pareliussen’s team found that seedlings planted near the edge of a remnant forest grew better than those planted further away. In contrast, shade structures did not improve seedling survival. In fact, one species grew worse in the shade.

Fig_TampoEdges

One vision for landscape-scale forest restoration on the Tampoketsa de Ankazobe. We used an edge-detection algorithm in Inkscape to highlight forest edges and valley bottoms, the places where trees grew best in our study. Imagery: Google Earth (2018).

Together, our two studies begin to suggest the outlines of a vision for landscape-scale forest restoration on the high Tampoketsa de Ankazobe. If native tree seedlings perform better along forest edges, it follows that a cost-effective strategy would be to focus on planting those areas first, leaving the higher, drier areas alone. Planting along edges would also be a conservative strategy given our hazy understanding of past landscapes. Some grasslands in highland Madagascar seem to be very old, and planting trees in such places could destroy habitat for grassland species, which are threatened in their own right.

For more information about this experiment, you can read our open access paper in Plant Diversity. We have also published several other blog posts about Ankafobe.

Hill of Honey: Forest Recovery on Madagascar’s Central Highlands

This post was co-written by Leighton Reid and Chris Birkinshaw after a three-day field trip in the tampoketsa with Cyprien Mandriamanana and Jeannie Raharimampionona.

A narrow, paved road winds north from Antananarivo through a high, windswept plain. It is the wet season, and the hills are green and close-cropped, but in the long dry season the landscape burns black. Orange rivers wind through the valleys, muddied by massive erosion. Here and there are thin strips of riparian forest, chock full of endemic species.

The biggest chunk of remaining forest is Ambohitantely, Malagasy for “hill of honey”. Ambohitantely encompasses 1,800 ha of humid forest (about three times the size of Saint Louis’s Forest Park). We visited the reserve to observe natural forest recovery in one of the few places it can still be seen on Madagascar’s Central Highlands.

Ambohitantely map

Ambohitantely: the last large tract of forest on Madagascar’s Central Highlands. Ten kilometers to the northwest is Ankafobe, a much smaller forest fragment managed by a local community with assistance from Missouri Botanical Garden.

Forest transition feedback

Why is forest recovery so rare in the Malagasy highlands? Madagascar’s Central Highlands are currently undergoing a complete ecological transition, from forest and wooded savanna to grassland. The degradation cycle often starts when people cut forest trees to extract wood for timber and charcoal production. Small-scale cutting opens the canopy, dries the forest floor, and creates debris, all of which increase forest vulnerability to annual wildfires that sweep across the grasslands during the eight month dry season. C4 grasses quickly colonize the burned land, inhibiting forest recovery, and creating ideal conditions for future fires. The reason that Ambohitantely has natural forest recovery for us to observe is because reserve staff maintain a wide fire break for more than thirty kilometers around the reserve.

Our guide at Ambohitantely took us on a hike through several areas where forest had once been cleared and burned but where fire had been excluded for 15 or 25 years, allowing the forest to begin to recover. Frankly, the vegetation was uninspiring. Low shrubs and forbs were scattered through a matrix of C4 grass (mostly Aristida), but trees and tree seedlings were nowhere to be seen outside of the forest (where they were abundant). By tropical forest standards, this looked slow. But for natural forest recovery on the Central Highlands, it’s hard to imagine a better situation than being protected from fire and immediately adjacent to the largest remaining tract of forest.

Ambohitantely forest edge

After 25 years of recovery, forest edges were sharp. Tree seedlings were almost totally restricted to forest, and grasses dominated the ground layer just outside.

Our main interest in Ambohitantely was to compare natural forest recovery there to our observations at another forest fragment, Ankafobe, 10 km northwest. For the last decade, MBG has partnered with a local community to preserve a thin, riparian forest containing several critically endangered plant species. Community members constructed a fire break and have begun to restore the surrounding hillsides by turning over the orange clay and planting fast-growing legumes to develop the soil.

Ankafobe hillside with Schizolaena

A degraded hillside at Ankafobe; the remnant forest is down the hill to the right, near the edge of the photo. The tree at the top of the hill is Schizolaena tampoketsana, a critically endangered microendemic in an endemic Malagasy plant family (Sarcolaenaceae). It is a remnant forest tree that likely escaped repeated fires by being nestled in a deep, protective gully. The multi-stemmed shrub in the foreground is actually a large tree species, Brexia montana, which has likely resprouted many times from a well-developed root system.

Some of our comparative observations were promising. We were happy to find examples at Ambohitantely where recovering land dominated by heather and blueberry seemed to have continued developing into a more diverse thicket, including Nuxia capitata, Psiadia altissima, and Razafimandimbisonia minor. At Ankafobe, we had worried that the heather growing in some areas signified poor soil conditions and possibly arrested development.

Overall our visit left us with more questions than answers. We hope to answer at least a couple of them over the coming years.

  • Why is natural forest recovery so slow on the Central Highlands?
  • Has it always been this slow?
  • Are Malagasy tree species poor pioneers because of their long, relatively stable evolutionary ecological history?
  • Is there any way to make Malagasy trees grow any faster on the degraded grasslands?1
  • Are the fire-stoking C4 grasses introduced from east Africa rather than being native species?
  • If so, when?
  • Is the soil too far gone to ever recover?
  • How important were now-extinct seed dispersers and grazers, like 200-kg lemurs and elephant birds?
Ambohitantely cloud cover

Although Ambohitantely is the only remnant forest fragment of any size, we learned recently that it may not be a perfect reference system for Ankafobe. For one thing, Ambohitantely is slightly higher and farther east, which results in considerably greater cloud cover during the dry season. This probably ameliorates the harsh conditions outside of the forest, at least a little. Shown here is cloud frequency from May – October, taken from NOAA MODIS satellite imagery. Thanks, Michael Douglas!

References

Goodman, S.M. & Jungers, W.L. 2014. Extinct Madagascar: Picturing the Island’s Past. University of Chicago Press, Chicago, IL.

Pareliussen, I., Olsson, E.G.A. & Armbruster, W.S. 2006. Factors Limiting the Survival of Native Tree Seedlings Used in Conservation Efforts at the Edges of Forest Fragments in Upland Madagascar. Restoration Ecology 14: 196-203.

1Two datasets (one from our team and one from Pareliussen et al’s [2006]) suggest that NPK fertilizer, even in relatively small doses, reduces native tree seedling performance. It is unclear whether this is because of toxicity (a direct effect) or because other plants, like shrubs, are better able to utilize the nutrient pulse and then compete more strongly against the native tree seedlings (an indirect effect).

A Tale of Two Highlands Part II: Ankafobe, Madagascar

Leighton Reid, James Aronson, and Chris Birkinshaw all contributed to this post on restoration in one of Missouri Botanical Garden’s community-based conservation sites in Madagascar. They are currently travelling together discussing opportunities for ecological restoration in MBG’s Madagascar Program and more generally for the country as a whole.

Madagascar’s central highlands appear as a grassy sea – an undulating terrain with intermittent red gashes where heavy rain has dramatically eroded the landscape. Driving north along the national highway from the capital, Antananarivo, one sees Eucalpytus trees growing near villages, as fuel and firewood plantations, but there is almost no natural forest. The few natural communities that remain represent vestiges of a former world.

The view across the road from Ankafobe - nearly unbroken grassland.

The view across the road from Ankafobe – nearly unbroken grassland.

Our destination today is one such vestige – the Ankafobe reserve. Ankafobe is a tiny (33 hectare) strip of native forest growing near the headwaters of a highland stream. Water-loving Pandanus trees demarcate the stream bed and provide fruits for several lemur species. A Souimanga Sunbird (Cinnyris sovimanga) flitters from tree to tree. Just outside of the forest, highly flammable grassland stretches to every horizon.

Pandanus spikes stand out in this thin patch of gallery forest at Ankafobe.

Fragmented gallery forest at Ankafobe. Spikey Pandanus demarcate the streambed. Red strips in the background are incipient forest restoration plots, where the soil has been turned over prior to planting nitrogen-fixing shrubs and native trees.

MBG staff and local villagers are working to restore forest on these bare hills, but it is not an easy task. Between clumps of grass is baked, orange laterite – rock hard soil bereft of life and nutrients. Tree seedlings planted in it grow slowly, or not at all. To improve seedling growth, MBG scientists are testing several strategies. One method is to turn over the soil and seed hearty legumes, whose symbiotic bacteria replenish soil nitrogen – a key ingredient in DNA.

Last October, a wildfire jumped the double fire breaks surrounding Ankafobe and burned a piece of the forest. Two hundred people from the local village (with a population of 600) voluntarily and spontaneously fought the fire for three days. Their impressive response minimized damage to this small forest and raised hopes and excitement about working together on conservation going forward.

The wildfire highlighted this forest fragment’s vulnerability, but it also provided a unique opportunity to observe the response to fire by a natural biotic community that has almost disappeared from the world. A number of trees were completely burned up that had been growing in the savannah just outside of the forest. Unexpectedly, several of these resprouted from their base and from superficial roots at some distance from the main stem. Nearby, the burned grassland bloomed an interesting  array of geophytic plants – particularly orchids – that were rarely observed in unburned grassland. These observations seem to support the hypothesis that at least part of the highland flora may be adapted to fire – a controversial idea that complicates the already challenging task of managing Ankafobe.

Ankafobe is a rare gem; a green emerald that stands out from the surrounding countryside and supports at least one species found almost nowhere else. The reserve is also a special opportunity for ecological restoration. Hard-won lessons from this site could eventually be used to restore tens of thousands of square miles of Madagascar’s central highlands.

Chris Birkinshaw (center) and the Ankafobe restoration team after a rainy afternoon in the field.

Chris Birkinshaw (center) and the Ankafobe restoration team after a rainy afternoon in the field.