Spatial models with covariates improve estimates of peat depth in blanket peatlands

Abstract

Peatlands are spatially heterogeneous ecosystems that develop due to a complex set of autogenic physical and biogeochemical processes and allogenic factors such as the climate and topography. They are significant stocks of global soil carbon, and therefore predicting the depth of peatlands is an important part of establishing an accurate assessment of their magnitude. Yet there have been few attempts to account for both internal and external processes when predicting the depth of peatlands. Using blanket peatlands in Great Britain as a case study, we compare a linear and geostatistical (spatial) model and several sets of covariates applicable for peatlands around the world that have developed over hilly or undulating terrain. We hypothesized that the spatial model would act as a proxy for the autogenic processes in peatlands that can mediate the accumulation of peat on plateaus or shallow slopes. Our findings show that the spatial model performs better than the linear model in all cases—root mean square errors (RMSE) are lower, and 95% prediction intervals are narrower. In support of our hypothesis, the spatial model also better predicts the deeper areas of peat, and we show that its predictive performance in areas of deep peat is dependent on depth observations being spatially autocorrelated. Where they are not, the spatial model performs only slightly better than the linear model. As a result, we recommend that practitioners carrying out depth surveys fully account for the variation of topographic features in prediction locations, and that sampling approach adopted enables observations to be spatially autocorrelated.

Mapping of Peat Thickness Using a Multi-Receiver Electromagnetic Induction Instrument

“…Considering environmental data, a few studies mostly used terrain attributes, including elevation and slope [2,3], only elevation [4] or the distance to a river [5]. More recently, machine learning models have also been used [6][7][8] as well as regression kriging [9]. Rudiyanto et al [6] (2016) and Young et al [9] (2018) built models using a relatively limited amount of environmental data (i.e., elevation, slope, aspect, System of Automated Geoscientific Analyses Wetness Index (SAGAWI) and nearest distance to river for the first study, and elevation, slope, aspect, vegetation type and soil for the latter).…”

“…More recently, machine learning models have also been used [6][7][8] as well as regression kriging [9]. Rudiyanto et al [6] (2016) and Young et al [9] (2018) built models using a relatively limited amount of environmental data (i.e., elevation, slope, aspect, System of Automated Geoscientific Analyses Wetness Index (SAGAWI) and nearest distance to river for the first study, and elevation, slope, aspect, vegetation type and soil for the latter). Rudiyanto et al [8] (2018) and Aitkenhead [7] (2017) used a wide array of environmental covariates: topography (i.e., elevation, vegetation-corrected elevation, and two derived terrain attributes-the Multi-Resolution Index of Valley Bottom Flatness (MRVBF) and SAGAWI-Euclidean distances to rivers, seas and combined rivers and seas, radar images (i.e., Sentinal-1A and ALOS-PALSAR) and vegetation (i.e., seven Landsat raw bands and the normalized difference vegetation index) in Rudiyanto et al [8] (2018), and topography (i.e., elevation and seven derived terrain attributes), climate (i.e., 24 different meteorological layers), soil (i.e., land cover, geology and soil maps) and vegetation (i.e., Landsat raw bands and derived vegetation indices) in Aitkenhead [7] (2017).…”

Estimating organic surface horizon depth for peat and peaty soils across a Scottish upland catchment using linear mixed models with topographic and geological covariates

“…A stratified sampling approach ensured that covariates were represented across their full ranges within the calibration and validation datasets (Fig. 2), in line with recent recommendations (Young et al, 2018).…”

“…Relationships between slope, elevation and organic layer depth have been used successfully to estimate blanket peat thickness in the Wicklow Mountains, Ireland (Holden & Connolly, 2011) and across land units with different soil and vegetation classifications in Dartmoor, south-west England (Parry et al, 2012;Young et al, 2018).…”

The Least Square Adjustment for Estimating the Tropical Peat Depth Using LiDAR Data

“…Several methods commonly applied for depth estimation are corpt, scorpan, trend surface, and kriging [20]. Other researchers also used univariate regression, multilinear regression, statistics, and linear regression methods [14,22]. Different techniques to estimate soil depth include multivariate geostatistics [16,22], inference modeling [23], and deep machine learning [18,24].…”