Header

UZH-Logo

Maintenance Infos

Spatio-temporal land use dynamics and soil organic carbon in Swiss agroecosystems


Stumpf, Felix; Keller, Armin; Schmidt, Karsten; Mayr, Andreas; Gubler, Andreas; Schaepman, Michael E (2018). Spatio-temporal land use dynamics and soil organic carbon in Swiss agroecosystems. Agriculture, Ecosystems & Environment, 258:129-142.

Abstract

Land use change strongly affects soil organic carbon (SOC), which is a critical driving force for soil quality and soil services. However, the influence of detailed land use dynamics on SOC remains an active research field. We investigated the impact of long-term grass-/cropland conversion dynamics on SOC for the Swiss agroecosystem (8‵500 km2). In Switzerland multi-temporal SOC observations are sparse, while the spatio-temporal resolution of available land use data is inadequate to capture small structured and dynamic land use. Thus, we stratified one-time SOC observations of two periods (n1995–1999 = 1096; n2011–2015 = 502) into typical land use classes of grass-/cropland conversion regimes to analyze states and trends of SOC. First, based on spectral imagery and auxiliary variables, we established a Random Forest classifier to detect annual grass-/cropland distributions in a spatial resolution of 30 × 30 m for 2000–2015. Second, we used the annual land use maps to derive classes of typical land use dynamics based on a pattern description of conversion regimes. Third, we comparatively assessed the SOC covariate importance (IMP) and SOC sample representation of land use dynamics using terrain and climate covariates. Subsequently, we statistically analyzed SOC across land use dynamics and two periods. The land use classifier shows an Overall Accuracy of 86%, while the annual land use maps deviate by 6%–11% from census data. 46% of the farmland underlies frequent grass-/cropland conversions. The SOC covariate importance of land use dynamics is increased by at least 2.4 times compared to the SOC covariates. The SOC observations consistently represent land use dynamics across SOC covariates. In both periods, SOC increased as the grassland share within the land use regimes increased. The mean SOC for permanent grasslands is 35 g kg−1 and 17 g kg−1 for permanent croplands. Areas of land use conversions where grassland is dominant show increased SOC (29–30 g kg−1) compared to areas where cropland was dominant (23–25 g kg−1). The mean differences of SOC between the periods and classes are statistically not significant except for permanent cropland, for which the SOC is decreased by 5.2 g kg−1.

Abstract

Land use change strongly affects soil organic carbon (SOC), which is a critical driving force for soil quality and soil services. However, the influence of detailed land use dynamics on SOC remains an active research field. We investigated the impact of long-term grass-/cropland conversion dynamics on SOC for the Swiss agroecosystem (8‵500 km2). In Switzerland multi-temporal SOC observations are sparse, while the spatio-temporal resolution of available land use data is inadequate to capture small structured and dynamic land use. Thus, we stratified one-time SOC observations of two periods (n1995–1999 = 1096; n2011–2015 = 502) into typical land use classes of grass-/cropland conversion regimes to analyze states and trends of SOC. First, based on spectral imagery and auxiliary variables, we established a Random Forest classifier to detect annual grass-/cropland distributions in a spatial resolution of 30 × 30 m for 2000–2015. Second, we used the annual land use maps to derive classes of typical land use dynamics based on a pattern description of conversion regimes. Third, we comparatively assessed the SOC covariate importance (IMP) and SOC sample representation of land use dynamics using terrain and climate covariates. Subsequently, we statistically analyzed SOC across land use dynamics and two periods. The land use classifier shows an Overall Accuracy of 86%, while the annual land use maps deviate by 6%–11% from census data. 46% of the farmland underlies frequent grass-/cropland conversions. The SOC covariate importance of land use dynamics is increased by at least 2.4 times compared to the SOC covariates. The SOC observations consistently represent land use dynamics across SOC covariates. In both periods, SOC increased as the grassland share within the land use regimes increased. The mean SOC for permanent grasslands is 35 g kg−1 and 17 g kg−1 for permanent croplands. Areas of land use conversions where grassland is dominant show increased SOC (29–30 g kg−1) compared to areas where cropland was dominant (23–25 g kg−1). The mean differences of SOC between the periods and classes are statistically not significant except for permanent cropland, for which the SOC is decreased by 5.2 g kg−1.

Statistics

Citations

Dimensions.ai Metrics
3 citations in Web of Science®
4 citations in Scopus®
Google Scholar™

Altmetrics

Downloads

2 downloads since deposited on 11 Jan 2019
2 downloads since 12 months
Detailed statistics

Additional indexing

Item Type:Journal Article, refereed, original work
Communities & Collections:07 Faculty of Science > Institute of Geography
Dewey Decimal Classification:910 Geography & travel
Uncontrolled Keywords:Agronomy and Crop Science, Ecology, Animal Science and Zoology
Language:English
Date:1 April 2018
Deposited On:11 Jan 2019 11:11
Last Modified:11 Jan 2019 11:12
Publisher:Elsevier
ISSN:0167-8809
OA Status:Closed
Publisher DOI:https://doi.org/10.1016/j.agee.2018.02.012

Download