In the Neotropics the establishment of pastures is associated with deforestation. A deeper understanding of the system indicates that farmers incorporate trees into their pasture management by managing live fences, dispersed trees in pastures, and forest fragments. Recent research indicates that the extent to which trees are incorporated into pasture management varies across farmers. Yet efforts to document and understand why we see this variation are scant. The main question arises: 1) does tree management vary across cattle pastures? If there are different tree management styles then we expect the frequency distributions of tree densities to show bimodality. Fifty four farms across 3 provinces in the Republic of Panama were visited in 2014. Sites were chosen to capture the different styles. Farm boundaries were obtained and Google Earth ® imagery was employed to digitize the tree canopy in each of the farms. The tree canopy was classified into landscape patterns based on the land-uses reported by farmers. GLM was employed to test if physical co-variates could explain farm tree density. Frequency distributions and density plots were generated for the total percent tree cover per farm and the percent area cover of each of the tree landscape patterns.
Results/Conclusions
The mean percent tree cover per farm area across 54 farms was 34% (SE = 2.12) ranging from a maximum of 83% to a minimum of 13.32 %. Findings indicate that percent tree cover can be categorized into seven landscape patterns: live fence (n = 54), dispersed trees in pasture (n = 53), riparian vegetation (n = 50), horticulture (n = 23), forest (n = 11), fallow (n = 11), forestry plantation (n = 10). Physical co-variates did not show an effect on tree density. Frequency distributions and density plots did not show strong evidence for bimodality across farms. Across 4 farms tree density was disproportionately higher which is attributed to the amount of fallow and dispersed trees in pasture. Clear bimodality was only evident for the live fences. Results demonstrate that farmers are employing different strategies as only few manage forestry plantations, forest fragments and fallow. Although there is no clear evidence of styles, four farmers are incorporating higher tree densities which suggests that with a larger sample size bimodality is possible. Given the importance of structural tree diversity for wildlife conservation we argue that tree management studies across cattle pastures are necessary for conservation and rangeland management.