Remote Sensing-Based Detection and Spatial Pattern Analysis for Geo-Ecological Niche Modeling of Tillandsia Spp. In the Atacama, Chile

Abstract

ABSTRACT:In the coastal Atacama Desert in Northern Chile plant growth is constrained to so-called 'fog oases' dominated by monospecific stands of the genus Tillandsia. Adapted to the hyperarid environmental conditions, these plants specialize on the foliar uptake of fog as main water and nutrient source. It is this characteristic that leads to distinctive macro-and micro-scale distribution patterns, reflecting complex geo-ecological gradients, mainly affected by the spatiotemporal occurrence of coastal fog respectively the South Pacific Stratocumulus clouds reaching inlands. The current work employs remote sensing, machine learning and spatial pattern/GIS analysis techniques to acquire detailed information on the presence and state of Tillandsia spp. in the Tarapacá region as a base to better understand the bioclimatic and topographic constraints determining the distribution patterns of Tillandsia spp. Spatial and spectral predictors extracted from WorldView-3 satellite data are used to map present Tillandsia vegetation in the Tarapaca region. Regression models on Vegetation Cover Fraction (VCF) are generated combining satellite-based as well as topographic variables and using aggregated high spatial resolution information on vegetation cover derived from UAV flight campaigns as a reference. The results are a first step towards mapping and modelling the topographic as well as bioclimatic factors explaining the spatial distribution patterns of Tillandsia fog oases in the Atacama, Chile.

Vegetation growth and landscape genetics of Tillandsia lomas at their dry limits in the Atacama Desert show fine‐scale response to environmental parameters

“…This results in dune growth and migration in an unusual upwind direction, and to optimize fog harvesting and minimize self-competition ("fog shadow"), the vegetation is structured in rows with defined space between them depending on slope. With increasing slope, the distance between rows is decreasing; however, there is an optimal slope ranging from between 5.5 and 12.5° because sand has to be trapped to stabilize growth (Wolf et al, 2016). Such linear or banded formation patterns in arid climates have already been described, but we lack a deeper understanding of growth processes (Aponte & Flores, 2013;Borthagaray et al, 2010;Deblauwe et al, 2012;Hesse, 2012Hesse, , 2014Tlidi et al, 2018), and to our knowledge, substantial biologically relevant data are not available (Koch et al, 2019).…”

“…All individuals were georeferenced according to high-resolution aerial photographs (Koch et al, 2019), and an accuracy of ~0.3 m was achieved. Accordingly, using an area model with detailed data on elevation at a minimum resolution of 30 cm (Wolf et al, 2016), positioning of individuals on a 100 × 100 m grid base is very precise. Genetic data of this sampling were previously reported (Koch et al, 2019), and herein, we discuss these data in the context of growth parameters.…”

“…Plants then function as a sand trap, and dune formation and migration faces the direction of the wind (Latorre et al, 2011). However, this system is optimally balanced with moderate slope (Figure 8, Wolf et al, 2016).…”

“…A positive feedback loop has not yet been observed. (b) Shows the environmental niche of T. landbeckii fields from the entire study area around Iquique in northern Chile (seeFigure 1; redrawn boxplots diagrams fromWolf et al, 2016)…”