Trends in active restoration of tropical dry forest: Methods, metrics, and outcomes
​
Introduction
​
Tropical dry forests (TDF), called seasonally dry tropical forests in some parts of the world, are the most endangered forest ecosystems in the world (Murphy and Lugo, et al., 2009). Compared to other forest ecosystems, TDF are understudied and lack adequate protection. Due to a lack of precipitation in TDF, high rates of fragmentation, and deforestation, TDF tend to regenerate slower (Janzen, D.H., 1988). Non-native plant species also out-compete native plant species, such as grasses which persist long after pastures are abandoned. While reviewing multiple case studies pertaining to restoration of TDF, many of the case studies were determined to be fundamentally inconsistent in terms of ecology. This is because the effects of different conservation methods are inconsistent in documenting or study, and outcomes of planting are vary highly. It was also found that economic and social were not thoroughly researched as many projects have not been successfully carried out or the true cost of implementing such methods have not been mentioned in most case studies (Dimson & Gillespie, 2020). Overall, this case study reviews the trends in TDF restoration outcomes.
​
​
Methods

For each article, there were attributes in five categories: publication details, site, characteristics, methods, metrics, and reported outcomes (see table 1). Papers were classified as restoration experiments or projects. Restoration experiments were empirical studies which evaluated the productiveness of active TDF restoration treatments. Projects were papers which documented results of active TDF restoration sites. Recorded information included location, conservation status, and cause of degradation. Plant attributes includes the number of individuals introduced and source native plant material. Sources were classified as “local” if they were “close to” or “nearby”, which is within 40 km of study area. Species attributes included the number of species planted in each study and reason for their selection (Dimson & Gillespie, 2020).
Seedling survival rates were compared among all studies to observe how different attributes affected seedling growth. Data from each study was included the study included quantitative data and the number of outplanted seedlings per population.
When searching for articles, Web of Science and Scopus data bases was utilized. Articles found were published between January 2000- June 2019. Key words used for search included: restor-, tropic- dry forest, dry forest, dry tropic, active, outplant, transplant, enrichment plant, and reintropduc-. Key words such as “reforestation” and “regeneration” were not used in order to direct the article search towards more ecological restoration centric peer reviewed papers (Dimson & Gillespie, 2020).

(Dimson & Gillespie, 2020)
Results
​
Sourced from Web of Science and Scopus, 60 articles were found, of which 30 met the specified criteria for the research review. The papers were published from 17 different journals, over half were published during or after 2013, and almost all papers were restoration experiments (see table 2).

(Dimson & Gillespie, 2020)
47% of studies were in US territories which included Hawaii, southern Florida, Puerto Rico, and US Virgin Islands. 20% of studies were in Mexico and the rest of the studies were in South America, Africa, and Asia. 60% of studies took place on protected areas, while the rest either occurred on private or university owned land (see figure 1). The most common reasons for land degradation were livestock grazing (50%), fires (37%), and cultivation (27%). Invasive species were also noted in 60% of the studies (Dimson & Gillespie, 2020). In regions such as Hawaii, approximately 90% of the land has been lost due to invasive species and deforestation for pastures (Gon & Olson, n.d.).
Monitoring Periods ranged from 75 days to 18 years, averaging 38.3 months. On average, restoration projects were longer than experiments. 63% of studies were ≤ 24 months long. Outplanting of native seedlings (71%) occurred more than seeding native plants (14%). Plant material was collected from local sources (43%) and on-site sources (32%). 91% of the studies focused on multi-species designs. Species richness was found to be higher in studies classified as restoration projects than experiments. Observations showed that plant species with functional traits were most common in species selection, followed by local abundance, and conservation goals. The least common reason for species selection was for local/ cultural significance, such as those which were rare, endangered, or sensitive to disturbance. Competition which would affect the experimental design was managed or assessed in 80% of the studies. This was done by active land management though herbicides, mowing/ trimming, or removal of fauna; in order to maintain the control variable (Dimson & Gillespie, 2020).
Metrics used to evaluate seedling growth included outplanted seedling performance and relative/ total increase in high or length (see figure 3). Variables regarding photosynthesis and plant stomata were noted as physiological variables. Native plant cover was noted as a community metric and species occurrence or abundance were noted as composition variables. Survival of outplanted seedlings ranged from 13%-80%, taken from 19,962 outplanted seedlings out of 105 treatments. There was a slightly positive correlation between outplant and number of outplanted species. A slightly lower positive correlation also existed between survival and length of monitoring. Results varied greatly with the various types of treatments. Survival of outplants were higher in protected sites with 54% chance, verses unprotected sites with a 30% chance of survival. Seedlings planted in heterogeneous assemblages had a slightly higher chance of survival at 63% then those planted in homogeneous assemblages at 29%. It was also observed that controlling non-native vegetation yielded a higher survival rate of 61% then not controlling non-native vegetation. Treatments which received irrigation throughout the study had a significantly higher survival rate at 63% then those that did not receive irrigation through out the study, at 51% (Dimson & Gillespie, 2020).

(Dimson & Gillespie, 2020)

(Dimson & Gillespie, 2020)

(Dimson & Gillespie, 2020)

(Dimson & Gillespie, 2020)
Discussion
​
Normally TDF restoration studies occurred within the US or Mexico, with most of them happening in Hawaii and lasted less than three years. Plant material for restoration was usually locally sourced and seedlings were preferred over seeds with mixed assemblages, instead of a single species of plant material. Studies usually took place in protected areas with domesticated or feral grazing livestock present, as most sites were abandoned pastures. Most papers focused on the rate of success of through individual plant performances to determine which methods were most effective.
​
The most common performance metric was survival of outplanted seedlings. Conventional forestry restoration usually favors seeds for plant reintroduction, but across most studies, it was shown that seedlings have a higher performance in restoration when compared to seeds. It was also found that outplanted plots of land had nearly twice the native plant cover than seeded plots, creating higher species richness in an area. It was also observed outplanted plots had relatively low survival rates due to low establishment of the plant in the environment as compared to seeded plots. Also noted that these trends may be due to high variation among treatments performed. Survival rates and quantitative data also vary in the amount of care and intensity in which the studies were conducted (Dimson & Gillespie, 2020).
​
It was observed that the survival rates metric overshadowed other important metrics such as reproduction, which is important for the long-term survival of TDF projects. Few studies noted positive performance results in terms of reproduction. One source observed higher reproductive rates for tall canopy plant species. Researchers speculate that the low recruitment rates for reproduction may be the low establishment of pollinators in the newly planted TDF, resulting in poor survival rates in such sites. This may also be the result of short monitoring periods which don’t allow newly established TDF to become properly settled in to produce more reliable data. The cost of the restoration projects and studies were also largely unreported, only mentioned in 20% of papers, as there were social and economic implications from each study. Land management, monitoring, and the undertaking of the projects took up approximately 98% of the costs and two additional years of land management represented 2% of the costs. The restoration projects being costly, volunteers were utilized in many of the projects (Dimson & Gillespie, 2020).
​
In order to create better research and practice studies, there needs to be more transparency in methods. For example, while focusing on one performance metric such as survivability, other metrics such as reproduction and costs are largely ignored in many studies. When producing new studies here onward, there have been many new tools at disposal for more accurate and rounded data such a remote sensing which would allow for greater monitoring ability and Thermal Radiometer Experiment Sensor (ECOSTRESS) which provides high resolution images from space depicting surface temperature in correlation to land stress and use. In addition, there have been many new data bases pertaining to land use and vegetation such as Vegetation Index (VDVI) which would give researchers more information to work with (Dimson & Gillespie, 2020).
​
Conclusion
​
The study concluded that most TDF restoration papers were empirical studies with few published long-term projects. Most studies utilized locally sourced seedlings in heterogeneous assemblages because of the higher success rates they yielded verses homogeneous seedling assemblages. It was also noted that land management of invasive species and herbivores was common among studies in order to maintain a controlled environment. In order to evaluate each study, metrics such as germination, growth, and survival were utilized to measure success of each treatment. The reviewed studies also lacked uniform strategies for how they carry out their studies as there was a high level of variation in layout and design of studies. They also failed to mention important metrics such as social and economic impacts of their studies in the area in the long run, the costs of conducting the project, and labor. Long term outcomes were also excluded from most studies (Wilson & Peter, 2016). In order for their to be a better understanding of TDF restoration, it is shown that there has to be more uniform and in depth project design as well as data collection; as well as including or providing better detail on under evaluated metrics such as social and economic impacts, and more long term studies (Dimson & Gillespie, 2020).
Citations:
​
Dimson, M., & Gillespie, T. W. (2020). Trends in active restoration of tropical dry forest: Methods, metrics, and outcomes. Forest Ecology and Management, 467, 118150. doi: 10.1016/j.foreco.2020.118150
Gon, S., & Olson, D. (n.d.). Hawaii tropical dry forests | Ecoregions | WWF. Retrieved from World Wildlife Fund website: https://www.worldwildlife.org/ecoregions/oc0202
Janzen, D.H., 1988. Tropical dry forest. In: Wilson, E.O. (Ed.), Biodiversity. National Academy Press, Washington, D.C., pp.130-137.
Murphy, P.G., Lugo, A.E., 1986. Ecology of tropical dry forest. Annu. Rev. Ecol. Syst. 17, 67–88. https://doi.org/10.1146/annurev.es.17.110186.000435.
Wilson, E. O., & Peter, F. M. (2016). Tropical Dry Forests The Most Endangered Major Tropical Ecosystem. Retrieved December 11, 2019, from Nih.gov website: https://www.ncbi.nlm.nih.gov/books/NBK219281/