Human disturbance on high Andean forests in Colombia: Spatio-temporal dynamics and vegetation analysis of forests fragments around Bogotá, and potential effects of climate change on treeline species distribution

Abstract

This research presents a multi-scalar work on the effects of human disturbance on high Andean forests in Colombia. It focuses on the analysis of vegetation composition and structure and on the dynamics of forest cover change and connectivity through time in selected localities around Bogotá. Additionally, a distribution model exercise is carried out to further the understanding of the environmental constraints driving the present distribution of a distinctive Andean treeline species and to examine its response to climate change in the future. High Andean forests are a unique and highly biodiverse ecosystem, located at high elevations in one of the world’s biodiversity hotspots, the Tropical Andes. Unfortunately, these forests are extremely fragmented and particularly vulnerable to climate change. Despite the fact that they provide essential ecosystem services to neighboring high Andean cities, such as the capital of Colombia, Bogotá, overall, high Andean forests have been poorly studied. Their fragility and their great value in terms of environmental services, together with their outstanding diversity, make them ecosystems with high priority for conservation. Nevertheless, especially around Bogotá, the vegetation studies carried out often have been rather circumscribed, rarely incorporating more complex parameters such as landscape fragmentation and anthropic disturbance metrics nor assessing functional and phylogenetic diversity attributes and ecosystem services. Also, there is a lack of studies that consider multiple vegetation strata at once and that address long-term changes in the extent and structure of high Andean forests, pairing plot-based work with forest cover change analysis. Lastly, few efforts have been made so far to model the response to climate change of high Andean forests and their species. Investigating these aspects could provide meaningful insights into species’ responses to anthropic influences and community assembly mechanisms, broadening our understanding of high Andean forests and strengthening strategies to conserve them. The aim of the present study is to explore the relationship between human disturbance and the condition of high Andean forest in the surroundings of Bogotá, the growing capital city of Colombia. Therefore, three parallel approaches were developed and implemented. On one hand, permanent plots establishment and plot-based fieldwork was carried out in several localities in the surroundings of Bogotá. Here, tree viii layer and understory vegetation were assessed and analyzed in terms of their taxonomic, functional and phylogenetic diversity, but also as to their structure and above-ground biomass production. A broad array of environmental factors and anthropic disturbance-related indicators was compiled and subsequently used to characterize the tree layer and understory plant communities using a cluster analysis in combination with ordination techniques (Non-metric Multidimensional Scaling and partial Redundancy Analysis). Finally, environmental and disturbance predictors were related to tree and understory diversity metrics and above-ground biomass by means of Generalized Linear Models (GLMs). As a result, a noticeable effect of anthropic related predictors on both tree and understory layer community composition and also a direct negative effect on tree layer diversity and above-ground biomass production have been documented. On the other hand, a thorough analysis of aerial pictures from the studied localities of the first study has been conducted, covering the period between the 1940s to early 2000s, in order to deepen the understanding of the history of forest fragmentation in the area. The images were classified into three land cover classes (Forest, Grasslands and Bare ground), utilizing texture analysis and object-based image segmentation. Then, the extent of forested areas was assessed through time. Additionally, forest connectivity metrics were calculated by performing a Morphological Spatial Pattern Analysis (MSPA) and correlated with the forest cover dynamics over the same period. For the period 1940-2000, in all except one of the studied localities a general restoration of forest cover and connectivity was determined. The highlighted forest cover and connectivity trends were related to the individual history of each locality in terms of population displacement and urban expansion. Lastly, to further expand the geographical and temporal coverage of this study, species distribution models were performed for Weinmannia fagaroides (Cunoniaceae), a tree species that was found to be distinctive of the higher elevation forest fragments that were studied in this project. This species is found throughout the Central and Northern Andes, and is a typical component of high elevation treelines, ecotonal zones that are believed to be highly vulnerable to climate change impacts. The main aim was to understand which are the focal drivers of the current distribution of W. fagaroides, and to hypothesize its response to climate change in the future. At the same time, a comparison of the efficacy of environmental predictors pertaining to three different climate datasets (Chelsa 1.2, Worldclim 1.4 and Worldclim 2) and of ix three different modelling algorithms (GLMs, Random Forests - RF, MaxEnt) was carried out for predicting the current distribution of W. fagaroides. The current species distribution was found to be driven mostly by one temperature, one precipitation and one soil-related variables. In regards to the performance of the three different climate datasets no significant difference was found, although Chelsa 1.2 retrieved slightly higher model evaluation scores. As to the algorithms, RF and MaxEnt retrieved the highest model evaluation scores. Moreover, the modeled distribution of W. fagaroides was predicted to decrease by 38% and to shift towards higher elevations in the period 2060-2080 for a high CO2 emission scenario. Considering that at present only around 22% of the modeled species distribution is situated in protected areas, an adaptation of currently implemented conservation efforts is strongly recommended to protect this and other fragile high Andean treeline species. Overall, this study contributes to the characterization of forest fragments in the vicinity of Bogotá and is an important addition for a better understanding of the effects of anthropic disturbances on high Andean forests and their species. Also, several methodological approaches are provided that may serve as a basis to better understand the situation of other species and ecosystems in similar regions.