bioRxiv2024
Costa Rica implemented the worlds first national-scale Payment for Ecosystem Service (PES) program in 1996 and now protects over 200,000 hectares. By distributing wealth towards local land-stewards, Costa Ricas program has helped to limit deforestation at a national scale, but the large-scale ecological implications have yet remained unclear. Here, we use a massive ecoacoustic dataset to evaluate how this program has impacted the ecological integrity of PES forests across the entire Nicoya Peninsula. At the times and frequencies that are indicative of native biological activity, we reveal dramatic increases in the integrity of PES soundscapes, relative to those of natural protected areas. Specifically, natural regeneration sites were 97.79% more acoustically similar to reference forests (absolute mean similarity of 0.539) than they were to disturbed pastures, while acoustic recovery of plantations lags behind (79.66%; 0.489). These findings are strongly suggestive of large-scale ecological recovery, constituting some of the most robust evidence to date that restoration initiatives can benefit biodiversity on large spatial scales. Study overviewCosta Ricas PES program pays landowners to encourage forest recovery and compensate them for opportunity costs. Most payments subsidize land conservation, in which participants allow existing forests to naturally regenerate. Some payments are also offered to produce timber through plantations, which are often monocultures of exotic tree species. Despite the programs importance to Costa Ricas conservation efforts, little is known about whether these forest systems are recovering their natural characteristics. To investigate the dynamics of the PES program, we recorded continuous 6-day soundscapes in 119 sites across the Nicoya Peninsula of Costa Rica (Supplementary Figure 1). Specifically, we characterized the soundscapes across 4 land-use types: (i) 19 reference pastures, (ii) 43 PES monoculture tree plantations, (iii) 39 PES natural regeneration sites, and (iv) 18 reference forests. Sites from each land-use type are distributed across the Nicoya Peninsulas climate and edaphic gradient, allowing us to capture substantial variation in ecological outcomes. We determined the areas of acoustic space where most animals vocalize and where ecological responses to recovery were most likely to be detected (Figure 1). We then identified how and to what extent the soundscapes of natural regeneration and plantation sites had changed over the last 27 years. We find evidence that naturally regenerating forests within the PES have recovered substantially when compared to reference forests, while plantation systems lag behind (Figure 3). Our findings reaffirm the importance of ecosystem conservation, while suggesting redistributive policy mechanisms can accelerate nature protection at scale. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=116 SRC="FIGDIR/small/610944v1_fig1.gif" ALT="Figure 1"> View larger version (47K): org.highwire.dtl.DTLVardef@1e23f8org.highwire.dtl.DTLVardef@cbac64org.highwire.dtl.DTLVardef@1a2daaborg.highwire.dtl.DTLVardef@c2157e_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOFigure 1:C_FLOATNO Tailoring the acoustic space to improve acoustic analyses. Soundscape analyses can be improved by identifying the times and frequencies associated with the dynamics of interest. Using manual annotations and modelling approaches we determined that differences in the activity of native fauna, shown here in black, were best understood within frequencies 1 - 9 kHz (on the y-axis) and the times listed along the x-axis in yellow. Outside of this region of acoustic space, differences between land-use types are better explained by climatic differences or other acoustic events (shown in grey). C_FIG O_FIG O_LINKSMALLFIG WIDTH=134 HEIGHT=200 SRC="FIGDIR/small/610944v1_fig3.gif" ALT="Figure 3"> View larger version (17K): org.highwire.dtl.DTLVardef@111505org.highwire.dtl.DTLVardef@17b36cborg.highwire.dtl.DTLVardef@4a5057org.highwire.dtl.DTLVardef@e08a29_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOFigure 3:C_FLOATNO Soundscape comparisons reveal substantial recovery in PES natural regeneration sites. (A) Scaled similarity values to reference forests are shown for the remaining three land-use types, higher values indicate more shared acoustic characteristics. Box plots show the quantile ranges and median values for each land-use type. Colored points represent mean similarities for all 10-minute time bins included in the analysis, colored by the time of period of the day, while outliers are plotted again in black. Brackets indicate the results of a Wilcoxon post-hoc statistical test. (B) 24-hour cycles of SPMN for each land-use type are shown for the frequency bands in which similarities between natural regeneration sites and reference forests were highest (3 - 4 and 6 - 7 kHz). C_FIG