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

Abstract:

Tropical forests are significant carbon sinks and their soils’ carbon storage potential is immense. However, little is known about the soil organic carbon (SOC) stocks of Tropical mountain areas whose complex soil-landscape and difficult accessibility pose a challenge to spatial analysis. The choice of methodology for spatial prediction is of high importance to improve the expected poor model results in case of low predictor-response correlations. Four aspects were considered to improve model performance in predicting SOC stocks of the organic layer of a tropical mountain forest landscape: Different spatial predictor settings, predictor selection strategies, various machine learning algorithms and model tuning. Five machine learning algorithms: random forests, artificial neural networks, multivariate adaptive regression splines, boosted regression trees and support vector machines were trained and tuned to predict SOC stocks from predictors derived from a digital elevation model and satellite image. Topographical predictors were calculated with a GIS search radius of 45 to 615 m. Finally, three predictor selection strategies were applied to the total set of 236 predictors. All machine learning algorithms—including the model tuning and predictor selection—were compared via five repetitions of a tenfold cross-validation. The boosted regression tree algorithm resulted in the overall best model. SOC stocks ranged between 0.2 to 17.7 kg m-2, displaying a huge variability with diffuse insolation and curvatures of different scale guiding the spatial pattern. Predictor selection and model tuning improved the models’ predictive performance in all five machine learning algorithms. The rather low number of selected predictors favours Forward compared to backward selection procedures. Choosing predictors due to their indiviual performance was vanquished by the two procedures which accounted for predictor interaction.

Abstract:

With respect to sampling for regression-based digital soil mapping (DSM), the above all aim is to ensure that the spatial variability of the soil is well-captured without introducing any bias, while the design remains feasible according to operational constraints such as accessibility, man power and cost. Representativeness of the sample concerning the population to be sampled needs to be guaranteed in any regression-based modelling approach. Four selected sampling designs were adapted to show that basically any design may be optimised to represent the n-dimensional predictor space of a particular area, while selecting points is only permitted from a small accessible sub-area or from outside the area. Sampling efficiency may be evaluated based on the representation of the predictor space. However, not only each predictor’s probability function but also the interaction between predictors may have to be considered, to select a representative sample. Instead of sampling a previously un-sampled area with limited accessibility, the four sampling designs may also be used to subsample an existing dataset and, thereby, optimise a suboptimal dataset based on the predictor space of the area which shall be mapped by DSM.