Environmental changes in biodiversity hotspot ecosystems of
South Ecuador: RESPonse and feedback effECTs (FOR2730)

Abstract:

Geographic Information Systems and Remote Sensing are important contributors to Sustainable Forestry Management Plans. Remote sensing techniques for image interpretation provides the means to extract valuable information that could be expensive and time-consuming to obtain through field observations (Franklin et al. 2001). Spatial Products derived from the interpretation of airborne and satellite borne images feed Geographic Information Systems to develop strategies and methodologies for resource management, harvest planning, fire management, map production, and model predictions. (Yusmah et al. 2015) This study has three important objectives: to test the feasibility of template matching for the identification of single pine tree crowns, to conduct a delineation of pine plantations using relational features and to evaluate how single tree crown size affects the accuracy of the proposed method. Templates of single trees were produced in the software eCognition Developer. The sampling process comprised the random selection of 3000 single pine trees in 7 different test sites (test sites were grouped in 3 categories according to the single tree sizes). A first rule set to detect single tree crowns was developed in eCognition Developer, using three different template groups (4, 8 and 16 templates). Through an analysis of variance, the number of single tree crowns detected was compared for the different template groups. Using a second rule set in eCognition, the template matching algorithm combined with relational, spectral and contextual information were applied to delineate pine plantation areas. An accuracy assessment was performed in the test sites for all thematic classes identified. Finally, an Analysis of Variance evaluated the influence of single tree crown size on the overall accuracy. Potential applications and improvements to the proposed methodology for single tree crown detection and plantation delineation are proposed at the end of the document.

Abstract:

Texture information from passive remote sensing images provides surrogates for habitat structure, which is relevant for modeling biodiversity across space and time and for developing effective ecological indicators. However, the applicability of this information might differ among taxa and diversity measures. We compared the ability of indicators developed from texture analysis of remotely sensed images to predict species richness and species turnover of six taxa (trees, pyraloid moths, geometrid moths, arctiinae moths, ants, and birds) in a megadiverse Andean mountain rainforest ecosystem. Partial least-squares regression models were fitted using 12 predictors that characterize the habitat and included three topographical metrics derived from a high-resolution digital elevation model and nine texture metrics derived from very high-resolution multi-spectral orthophotos. We calculated image textures derived from mean, correlation, and entropy statistics within a relatively broad moving window (102 m × 102 m) of the near infra-red band and two vegetation indices. The model performances of species richness were taxon dependent, with the lowest predictive power for arctiinae moths (4%) and the highest for ants (78%). Topographical metrics sufficiently modeled species richness of pyraloid moths and ants, while models for species richness of trees, geometrid moths, and birds benefited from texture metrics. When more complexity was added to the model such as additional texture statistics calculated from a smaller moving window (18 m × 18 m), the predictive power for trees and birds increased significantly from 12% to 22% and 13% to 27%, respectively. Gradients of species turnover, assessed by non-metric two-dimensional scaling (NMDS) of Bray-Curtis dissimilarities, allowed the construction of models with far higher predictability than species richness across all taxonomic groups, with predictability for the first response variable of species turnover ranging from 64% (birds) to 98% (trees) of the explained change in species composition, and predictability for the second response variable of species turnover ranging from 33% (trees) to 74% (pyraloid moths). The two NMDS axes effectively separated compositional change along the elevational gradient, explained by a combination of elevation and texture metrics, from more subtle, local changes in habitat structure surrogated by varying combinations of texture metrics. The application of indicators arising from texture analysis of remote sensing images differed among taxa and diversity measures. However, these habitat indicators improved predictions of species diversity measures of most taxa, and therefore, we highly recommend their use in biodiversity research.