Ecological and spatial modeling: mapping ecosystems, landscape changes, and plant species distribution in Llanos del Orinoco, Venezuela

The transformation of Llanos del Orinoco, focused on the flooding savanna, is evaluated in terms of the change and replacement of the savanna ecosystem and the plant species distribution under a Landscape Ecological approach. This research is carried out at three spatial scales: sub-continental, regional and local. At sub-continental scale, monthly composites NOAA-AVHRR normalized vegetation index (NDVI) images were interpreted based on the phenology of savanna ecosystems in the Llanos del Orinoco region in order to produce an ecosystem map of the area.

At regional scale, an analysis and description of the landscape changes derived from the construction of dikes to control inundation in the flooding savanna of the Llanos del Orinoco in Venezuela is presented. Elaborations of the landscape ecological maps were based on a land unit approach. An analysis of the landscape transformation derived from the embankment of the flooding savanna is carried out based on the savanna ecosystems replacement, Besides, the hydrological dynamics of the flooding savanna landscape was analyzed through the Digital Elevation Model (DEM) of the El Frio Biological Station as a representative area of flooding savanna. Field measurements were carried out with real time kinematic geographical positioning system and air photographs as well as radar images are used for further interpretation and modelling. The results were a precise digital elevation model, showing distribution of sinks and insight in the process of filling the area in periods of flooding. This DEM was used to obtain a Relative Altitude Model of the study area, and integrated to the soil water relative content field observation in order to produce single spatial models of plant species distribution in a changing flooding savanna landscape by effect of the embankment in the Llanos del Orinoco.

At local scale, the direct and indirect response of vegetation to environmental factors of flooding savanna in the El Frio Biological Station area were analyzed using Canonical Correspondence Analysis (CCA) as direct gradient analysis technique and Path Analysis as indirect statistical technique. Data contain the frequency and cover percentage of plant species, physical and management environmental variables, remote sensed variables and mapping data from 37 sites. The first ordination axis was mainly associated to relative soil water content and the second ordination axis was related to grazing intensity. The hydrological dynamics and the capacity to accumulate water in the soil mainly determined the distribution of species in flooding savanna. Based on environmental-vegetation relationship, Gaussian logistic models were used to evaluate the response of plant species vegetation on environmental gradients of the area. The vegetation data collected contain information about the frequency and percentage of cover for 57 sampling units (sites). The result is that Leersia hexandra and Panicum laxum, which are the more abundant species, occupy different position into the environmental gradients. These results define the Ecological Model of plant species distribution.

Finally, the analysis and integration of ecological processes into multiple ecological-spatial models were carried out in order to understand, evaluate and predict the distribution of dominant plant species in a changing flooding savanna landscape by effect of the embankment in the Llanos del Orinoco, Venezuela. The models using multiple spatial variables presented best accuracy than mode!s of single variables. The application of ecological integrated model of several spatial variables could be useful to analyze possible changes in the plant species composition on local, regional or global change scenarios. Based on the concepts and main results described and analyzed in this Thesis, an integrated model of ecosystems distribution in the flooding savanna was assembled. This mode! is based mainly on the spatial distribution of the dominant species and a state and transition ecosystem model.