New paper: quantifying tropical reforestation rates using Landsat time series

How to restore tropical forest cover to degraded landscapes is a challenge at the forefront of environmental policy. Achieving international commitments to bring hundreds of millions of hectares of land under restoration will require strategic planning, including identifying sites where rates of forest succession are sufficient to restore tree cover with minimal intervention. However, our capacity to produce ecological forecasts of forest succession at landscape scales remains limited. In a new paper, we address this gap with models for forest succession that can be parameterized with satellite remote sensing data. While Landsat data has long been used to monitor deforestation rates, monitoring the continuous changes that occur during succession is more challenging, in part because satellite data contains many sources of noise unrelated to forest structural change. We present a state-space modeling framework that explicitly disentangles measurement error in Landsat-derived spectral reflectance from successional changes related to forest structure.

I’m particularly excited about the modeling efforts in this paper. State-space models are widely used in wildlife, epidemiology, and other applications with “messy” data, but have not yet been widely used for satellite imagery. We demonstrate that state-space approaches are a feasible technique for disentangling process and measurement error from time series of imagery. All of our code is available at:

Citation: Caughlin, T.T., Alvarez‐Buylla, C.B., Asner, G.P., Glenn, N.F., Bohlman, S.A., Wilson, C.H., In Press. Monitoring tropical forest succession at landscape scales despite uncertainty in Landsat time series. Ecological Applications, e2208.

Anthropogenic impacts of restoration and wildfire on sagebrush distribution

Historically, vast areas of cold desert in the Western United States were defined by the abundance of one keystone plant species: sagebrush. The iconic sagebrush plant with its aromatic greenish-gray leaves and gnarled woody trunk provides a range of ecosystem services. For example, sagebrush is the main food source for the West’s unique wildlife species, including sage grouse and pygmy rabbits.

The big sagebrush (left) dominates vast stretches of iconic landscapes in Western North America (right). Photo by Dr. Trevor Caughlin.

Despite their importance, sagebrush plants are disappearing from western landscapes, threatened by wildfires, invasive species, and urban growth. To curb the degradation of this imperiled ecosystem, land managers have spent hundreds of millions of dollars to restore sagebrush populations across their historic range in the Great Basin. The threats to sagebrush ecosystems raise the question of how sensitive sagebrush cover is to human impacts, both positive and negative. Answering this question is complicated by the enormous range of environmental conditions in the Great Basin, from Badwater Basin, the lowest point in North America, to Mount Whitney, the tallest mountain in the contiguous United States. Across the range of natural variation in the Great Basin, how important are human impacts for sagebrush?

Our new paper addresess the question of how human impacts have altered the distribution of sagebrush plants across >200,000 square miles of western land. In collaboration with agency partners at the United States Geological Survey, we assembled a record of data on sagebrush plants, fire history, and restoration treatments spanning nearly 30 years. We then applied mapping and mathematical models to tease apart the importance of wildfire and restoration from climate and topography. We found that wildfire history had negative impacts on sagebrush comparable to the effects of geographic variation in elevation, rainfall, and temperature. This result emphasizes the threat that changing wildfire frequency pose to ecosystem health in the Western United States.

At the same time, we found cause for hope. Restoration action can have a positive impact on sagebrush ecosystems. This was a surprising result, because the 35-year record of restoration treatments included a very wide range of management actions, from planting seedlings to invasive species removal. Altogether, these results point the way for models and mapping to inform conservation of an imperiled habitat.

The modeling approach provided by the researchers is applicable across a range of ecosystems and species. As human impact on nature grows, land managers need reliable methods to guide species conservation and ecological restoration in the face of global change threats. For example, fire frequency is increasingly altered by environmental change in the Great Basin, where millions of acres of public rangelands have been actively managed over the last 75 years. Similarly, a number of international initiatives have set targets to restore ecosystems to hundreds of millions of acres of degraded land worldwide. We demonstrate how documenting decades of land management history can help scientists understand human impacts on ecosystems.

Requena‐Mullor, J. M., K. C. Maguire, D. J. Shinneman, and T. T. Caughlin. (In Press). Integrating anthropogenic factors into regional-scale species distribution models — a novel application in the imperiled sagebrush biome. Global Change Biology.

Biocultural legacies in urban trees of a tropical city

NEW PAPER: Tree species distributions in tropical cities are shaped by many factors, including colonial history and cultural legacies. Our new paper, led by Nadia Hunte, quantified drivers of urban trees in Georgetown Guyana, South America’s only English-speaking country. The urban trees of Georgetown include several fruit trees that represent the unique cultural heritage of the city, including several species not recorded in cities of neighboring Brazil and Venezuela. These fruit trees include Katahar (a; Artocarpus camansi), a tree likely brought to Guyana by indentured servants from India, and Ackee (b; Blighia sapida) a tree associated with the slave trade from West Africa. Fruit trees in cities reflect unique biocultural legacies and have the potential to promote food security for urban populations. For more details see:

Citation: Hunte, N., A. Roopsind, A. A. Ansari, and T. Trevor Caughlin. 2019. Colonial history impacts urban tree species distribution in a tropical city. Urban Forestry & Urban Greening 41:313–322.

Demographic costs and benefits of natural regeneration for tropical trees

Whether to plant trees or enable forest to return naturally is a fundamental question for tropical forest restoration. An overlooked aspect of this debate is that both restoration strategies rely on natural recruitment to regenerate biodiversity, as native tree seedlings recruit in the understory of restoration sites. While we know a lot about the rates of recruitment during tropical forest restoration, the long-term fate of these seedlings has remained a mystery. This is problematic, because the long-term success of any restoration project relies on tree seedling completing their life cycle to become seed sources.  In a new paper, we track the fate of natural recruits over a 7 year period in experimental reforestation plots in Los Tuxtlas, Mexico. We quantified growth, survival, and reproduction rates and then used a computer model to combine all these vital rates and ask, what is the probability that a newly-arrived seedling can grow and survive to reach reproductive maturity? We found an overall benefit to natural regeneration: tree seedlings have a much higher chance of achieving reproductive status in these plots. Our work is one of the first full life-cycle analyses of trees during restoration and adds a new perspective on the ongoing debate between active restoration and natural regeneration.



A research agenda to predict restoration outcomes

Ecological restoration can be expensive and unpredictable. In a new paper, led by Lars Brudvig, we outline a research agenda to advance the predictive capacity of restoration. I’m really happy with how this paper turned out. In the next few years, I plan to carry out some of the modeling ideas we describe.


Reduced impact logging could benefit wildlife in tropical forests

Overhunting is degrading the ecological integrity of tropical forests across the globe. I’m pleased to be a co-author on a new paper led by Anand Roopsind, that points to solutions for this bushmeat crisis. Our study quantifies the impacts of reduced-impact logging and indigenous hunting on wildlife occupancy in the Iwokrama rainforest. We found evidence suggesting that wildlife populations were persisting in Iwokrama, a multiple-use forest with some logging and hunting. Iwokrama is fairly unique as a site that  (1) is certified as being responsibly managed for timber production and (2) legally guarantees indigenous communities hunting rights in the forest, while prohibiting hunting by outsiders. Our results provide hope that well-managed logging and hunting could support local livelihoods and conserve biodiversity.

Long time series of satellite data predict soil carbon in pasture

The publicly-available, multi-decadal Landsat archive is an amazing resource. In our new paper, led by Chris Wilson, we show that a 28-year record of Landsat-derived greenness can predict soil carbon in Florida cattle pasture. This work opens up many interesting ecological questions about how vegetation dynamics lead to soil carbon sequestration, in addition to potential applications for spatially-targeted land management.

Check it out:

Wilson, C.C., T.T. Caughlin, S.W. Rifai, E,H. Boughton, M.C. Mack, L.S. Flory. Multi-decadal time series of remotely sensed vegetation improves prediction of soil carbon in a subtropical grassland. Ecological Applications. In Press.


Our new paper made the cover in Ecological Applications!


Our study made the cover of Ecological Applications! The image shows tree crowns and landscape features from a hyperspectral aerial image in Panama’s Azuero peninsula.

In the new study, we show how to predict tropical tree growth rates from aerial flyover data. This is like a blood test for trees–a single data-rich measurement that provides insight into future health and performance. Our hope is that this research opens the door for “precision forestry,” including spatially-targeted interventions to promote reforestation. Lots of work went into the paper, from collecting field data in Panama to months of crunching numbers on the computer. So I’m happy to see it out!

A new model to scale up forest restoration from sites to landscapes

Restoring forest to hundreds of millions of hectares of degraded land has become a centerpiece of international plans to sequester carbon and conserve biodiversity. To achieve these ambitious restoration goals, we will need to predict restoration outcomes at landscape and regional scales. However, ecological field studies reveal widely divergent forest recovery rates, challenging our ability to predict reforestation across sites. Mathematical models present an opportunity to synthesize results from different studies for a general understanding of reforestation dynamics. In our new study, we develop a theoretical framework to ask how tree canopy closure, a critical turning point for secondary forest succession, depends on landscape features beyond the scale of most field sites. Our modeling framework predicts the dynamics of reforestation using parameters that are commonly estimated in field studies. You can explore our basic model using the sliders below:

The simple version of our model (above) shows how different survival, growth and seed arrival rates can lead to differences in tree canopy closure within a single site. The model can also be applied to ask how landscape features impact reforestation rate across multiple sites. In particular, seed rain into deforested sites is critical for forest recovery and depends on landscape features that are difficult to measure or replicate. For example, fruit bats, an important seed disperser in degraded habitats of Southeast Asia, can range over tens of kilometers in a single night—an area far larger than most field plots.

Field studies reveal seemingly-contradictory results on the importance of seed rain for reforestation. Some studies find that adding seeds to degraded sites increases seedling abundance, whereas others do not. Similarly, some studies find that landscape features that increase seed availability (such as amount of surrounding forest cover) increase woody stem diversity, whereas others do find these relationships. Our results explain why field studies in sites with different landscape configurations could find different effects of seed limitation on reforestation rate. We predict that in landscapes with either very low seed availability, such as abandoned sugar cane plantations, or very high seed availability, such as primary forest, landscape-scale features will have a minimal impact on reforestation rate. In landscapes with intermediate seed availability, where there is enough seed rain to initiate reforestation yet not enough to guarantee rapid canopy closure, we show that seed availability can have significant impacts on reforestation rate. These results demonstrate how landscape features can lead to divergent forest recovery between otherwise similar patches.

Our new paper illustrates how mathematical models can provide conceptual insight into scaling up tropical forest restoration from sites to landscapes. More realistic models will be required to guide the spatial planning of specific forest landscape restoration projects. Adding realistic but complicating factors to theoretical models, such as fire disturbance and tree species interactions, is an exciting research frontier for restoration ecology. We anticipate that future modeling efforts will require collaboration between field ecologists and model developers to navigate between model realism and mathematical complexity. Ultimately, translating quantitative forecasts into spatially-targeted interventions for forest landscape restoration could help achieve ambitious goals of restoring hundreds of millions of hectares of tropical forest to degraded land.

Image credits: Landscape photo by Trevor Caughlin. Yep, the equation in the background is the primary model in the new paper!

Caughlin, T. T., S. Elliott, and J. W. Lichstein. In Press. When does seed limitation matter for scaling up reforestation from patches to landscapes? Ecological Applications. DOI: doi:10.1002/eap.1410

New paper: Loss of animal seed dispersal increases extinction risk in a tropical tree species

collecting civet seeds off a log

Happy to announce a new publication:

Caughlin T.T., Ferguson J.M., Zuidema, P.A., Levey, D.J., Bunyavejchewin S.,Lichstein J.W. Loss of animal seed dispersal increases extinction risk in a tropical tree species due to pervasive negative density dependence across life stages.In press, Proceedings of the Royal Society B: Biological Sciences

Overhunting directly threatens mammals in tropical forests worldwide and could indirectly threaten trees with seeds dispersed by mammals. Without seed dispersal, seeds remain crowded beneath the parent tree. Using field data and simulation models, we investigated the long-term effects of seed dispersal for a tree species in Thailand. We found negative effects of crowding for growth and survival across the entire tree life cycle, from seeds to adults. Loss of mammalian seed dispersal increased crowding and raised the risk of tree extinction by ten-fold. Our findings suggest that overhunting could lead to cascading extinctions in tropical forests.

This paper is a result of my dissertation research, including a ton of field work (the above photo is me excitedly poking through civet poop during the first year of the project) and a ton of computer programming. So I’m glad to see it finally out in print.




New postdoctoral fellowship!

Excited to announce that I have received an NSF postdoctoral fellowship to research landowner decision-making and landscape-level reforestation under the Science, Engineering and Education for Sustainability (SEES) Fellows program. During the grant, I will be based at the UF School of Forest Resources and Conservation, with Stephanie Bohlman as research mentor, and Dan Brown (U Michigan) as co-mentor.