The potential of secondary forests to restore biodiversity of the lost forests in semi-deciduous West Africa

Abstract

In West Africa, more than 80% of the original forest cover has disappeared due to the exponential growth of human populations in a recurrent search for new agricultural land. Once the fertility of the land is exhausted, these areas are abandoned and left to be reforested through natural succession. Despite the widespread presence of secondary forests of various ages in West African landscapes, little is known about the trajectories of recovery and the environmental factors that influence recovery rates. We set up 96 0.2 ha forest plots, along a chronosequence of 1 to 40 years and including 7 controls, on which all trees larger than 2.5 cm in diameter at breast height were inventoried. We modelled the recovery trajectories of four complementary dimensions of biodiversity (richness, diversity, composition, indicators of old-growth forest) in a Bayesian framework. Our results show that the four dimensions of biodiversity recover at different rates, with composition recovering much faster than floristic diversity. Among the local, landscape, and historical factors studied, the number of remnants and proximity to old-growth forests have a positive impact on recovery rates, with, under good environmental conditions, the composition, richness, and diversity being almost completely recovered in less than 25 years. Our results demonstrate the very high resilience of the composition of the semi-deciduous forests of West Africa, but also suggest that the management of these post-forest areas must be differentiated according to the landscape context and the presence of isolated trees, which are the last vestiges of the former forest. In unfavourable conditions, natural dynamics should be assisted by agroforestry practices and local tree planting to allow for a rapid restoration of forest goods and services to local populations.

Introduction

Global forest cover is estimated at around 4 billion hectares (FAO, 2020) and tropical forests contain most of the tree diversity of this forest area with around 53,000 tree species (Slik et al., 2015). Unfortunately, during the previous decade (2000−2010), the annual area deforested was about 10 million hectares (FAO, 2020) and this was mainly caused by the strong growth of the world's population which led to high demand for agricultural products (Crist et al., 2017). This situation, which is still ongoing, continues to favor the conversion of forest land to agricultural land. In West Africa, the human population growth rate has been estimated at 2.7% per annum over the last 4 decades, resulting in de facto high food needs and intense forest disturbance by a population still in search of fertile land (Brandt et al., 2017). The direct consequence is that more than 80% of the original forest cover has disappeared in West Africa and residual forests continue to be under high human pressure (Amahowe et al., 2018). Yet, the tropical rainforests of West Africa are among the most original forests in the world in terms of the diversity of species (Myers et al., 2000) even though they are now considered the most threatened (Aleman et al., 2017).

Faced with the continuous decline in primary forest cover, the restoration of secondary forests (i.e. old fields or abandoned pastures that are naturally recovering) has become a global issue. Initial empirical works quickly suggested that tropical forests are generally very resilient after a period of cultivation (Chazdon, 2003) provided that natural regeneration processes are not interrupted (Arroyo-Rodríguez et al., 2015). Thus, after the abandonment of cultivated land, secondary forests can rapidly recover acceptable levels of biodiversity (Chazdon et al., 2009), and this potential for species accumulation could constitute a capital reservoir for biodiversity conservation (Ferraz et al., 2014; Gilroy et al., 2014) especially in areas where primary forests have been severely degraded and/or have disappeared. Most of the work has been carried out in Central and South America where secondary forests would be able to resemble old-growth forests after a few decades of succession (Chazdon et al., 2009; do Nascimento et al., 2012), although ecosystem trajectories are not very easy to predict (Norden et al., 2015). Part of the difficulty of prediction stems from the diversity of environments in which secondary forests develop. Indeed, several studies have shown that certain local (e.g. soil type), landscape (e.g. forest context), or historical (e.g. the number of years of cultivation) factors can greatly vary the return rate of typical forest-dwelling species. Overall, there is a beneficial effect of remnant trees and forest proximity in the recovery of diversity in secondary forests (Duarte et al., 2010). Remnant trees modify the microclimate and attract seed-dispersing animals such as birds and bats, thus facilitating recolonization after the abandonment of fields (Derroire et al., 2016). The presence of an old forest nearby provides a source of propagules for the secondary forest, which also allows it to recover faster and with high species diversity (Meiners et al., 2015). Continental-scale synthesis suggested that it takes 20 years only in South America for a secondary forest to recover 80% of the species richness of an old forest (Rozendaal et al., 2019).

In contrast to the Neotropics, knowledge of the dynamics of secondary forests remains generally very rudimentary in Africa, particularly in West Africa (Norris et al., 2010). Numerous studies have been carried out on the vegetation found in secondary forests (N'guessan and N'Dja, 2018; Kassi et al., 2017). However, the latter often remain very descriptive, either by studying plant communities from a botanical and/or biogeographic point of view (N'Dja and Decocq, 2008), or by relating the evolution of vegetation structure and its carbon recovery dynamics (N'Guessan et al., 2019). To our knowledge, there is almost no work dedicated to the modeling of biodiversity recovery in the secondary forests of West Africa (Norris et al., 2010). Even fewer results are available on the factors that may or may not influence these dynamics of recovery of biodiversity. At this stage, It is important to stress that biodiversity is not a simple concept and that several dimensions are necessary to address it in its different facets (Willis et al., 2005). The most common and widely used measure used by managers of forested areas is species richness (Chaudhary et al., 2016), i.e. the number of species encountered per unit area which is also called diversity of order 0. However, behind this specific richness, very different distributions of relative abundance can be hidden and to capture them, it is interesting to complement the measure of richness with a measure of equitability, as defined by order 2 diversity, better known as Simpson's diversity (Marcon et al., 2014). This is all the more relevant since poor equitability is often observed in the early stages of forest development, stages dominated by one or a few pioneer species, and this equitability then returns more or less rapidly through complementarity and then selection effects (Schmitt et al., 2020). Furthermore, neither specific richness nor the diversity of order 2 gives us interesting information on the composition of communities, and notably on its composition in relation to the composition of old-growth forests, often considered as reference forests at the end of the trajectories of recovery (Mirabel et al., 2020). It would therefore seem appropriate to measure the level of community recovery also by measuring similarity in composition with old-growth forests. Finally, among all the species that shape the composition of old-growth forests, certain species, known as indicator species, are of particular interest because they are the species (i) that are the most dependent on the microclimate of the forest ecosystem and (ii) that are also often of great heritage interest since they are typically the first to be lost in degraded or secondarised forests (Björklund et al., 2020).

We have therefore chosen to use the different dimensions (richness, diversity, composition, indicators) of biodiversity to describe the different recovery trajectories of secondary forests in a post-forested landscape in West Africa. We wanted to identify the main local, landscape, and historical determinants of the recovery and to draw lessons for the management of these forests in the perspective of a possible role of secondary forests in the conservation of African biodiversity (Norris et al., 2010). To this end, we set up 89 plots in secondary forests and 7 plots in old-growth forests, each with an area of 0.2 ha, along a chronosequence of 1 to 40 years and on which we inventoried and determined all trees over 2.5 cm in diameter at breast height. We chose to carry out this study within a single forest massif, located in the semi-deciduous belt of West Africa, because we wanted to reduce climatic variations, which are very rapid on the North-South gradient in West Africa (Barry et al., 2018), and phytogeographic variations, which would complicate the construction of reference levels for compositional trajectories (Droissart et al., 2018). In this paper, we ask the following questions:

• What are the recovery trajectories of the different facets of biodiversity (specific richness, diversity, composition, indicators of old-growth forests)?

• What are the relative effects of local, landscape, and historical factors on the recovery rates?

• What are the consequences for the effective management of these secondary forests with a view to restoring the forest ecosystems of West Africa?