Effects of initial functional-group composition on assembly trajectory in savanna restoration

Prepared by André G. Coutinho, Monique Alves, Alexandre B. Sampaio, Isabel B. Schmidt & Daniel L. M. Vieira

Results of active restoration in Chapada dos Veadeiros National Park, Goiás, Brazil (14°7’2.54″ S 47°38’30.36″ W). The area on the left side (A) was restored by direct seeding, leading to a vegetation structure more similar to a native Cerrado. Two native species can be seen in this area: Lepidaploa aurea, a short-lived shrub, and Aristida riparia, a perennial grass. The area on the right side (B) was not restored and is dominated by the invasive grass Urochloa decumbens. Photo credit: Alexandre Bonesso Sampaio.

Active restoration of savanna ecosystems is more recent compared with forest ecosystems, and, in many cases, it is still difficult to predict if an area under restoration will reach the desired conditions. The final community of species may depend upon several factors: the restoration method used, the abiotic conditions, how many and which species are introduced, and the presence of native and invasive species within or in the surrounding of the restoration area. In our study, we monitored the restoration of an area of Cerrado, the Brazilian savanna, by direct seeding, an active restoration method that has been applied recently for restoring grasslands and savannas. We observed the changes of the proportion of species during two years and verified that the community after this period is affected by the dominant functional group (the characteristics of the dominant species) at the beginning of the monitoring. Species that are short-lived but grows quickly and can cover the ground in a short time were important for the establishment of perennial grasses in the following two years, possibly because they prevented soil erosion and maintained good soil conditions. On the other hand, invasive grasses increased a lot in the restoration site, displacing native species in some localities within the area. This happened because the control of invasive species applied before seeding was not efficient enough to deplete the seed bank of invasive grasses, which were able to re-establish in the restored area.

Our results indicate the importance of introducing different functional groups when restoring severely degraded savannas, where the resprouting ability has been lost. A mixture of fast-growing/short-lived species with perennial species increases the chances of an establishment of native species through a fast species turnover. The inclusion of seeds of trees is also important to create the tree layer typical of savannas. We also recommend a strong control of invasive species, which can be done through the combination of different methods: usage of herbicides, prescribed fire, plowing the soil or topsoil removal. Considering these recommendations should increase the success of restoration strategies in severely degraded savannas.

This is a plain language summary​ for the paper of André G. Coutinho and colleagues, published in the Applied Vegetation Science (https://doi.org/10.1111/avsc.12420).