Assessing the reliability of peatland GPP measurements by remote sensing: From plot to landscape scale

Using machine learning algorithms to predict groundwater levels in Indonesian tropical peatlands

Tropical peatlands play a vital role in the global carbon cycle as large carbon reservoirs and substantial carbon sinks. Indonesia possesses the largest share (65 %) of tropical peat carbon, equal to 57.4 Gt C. Human perturbations such as extensive logging, deforestation and canalization exacerbate water losses, especially during dry seasons, when low precipitation and high evapotranspiration rates combine with the increased drainage to lower groundwater levels. Drying and increasing temperatures of the surface peat exacerbate ignition and wildfire risks within the peat soils. As such, it is critically important to know how groundwater levels in peatlands are changing over space and time. In this study, a multilinear regression model as well as two machine learning algorithms, random forest and extreme gradient boosting, were used to model groundwater level over the study period (2010-12) within a peat dome impacted by drainage canals and multiple wildfires in Central Kalimantan, Indonesia. Although all three models performed well, based on overall fit, spatial modeling of groundwater level results revealed that extreme gradient boosting (R² = 0.998, RMSE = 0.048 m) outperformed random forest (R² = 0.997, RMSE = 0.054 m) and multilinear regression (R² = 0.970, RMSE = 0.221 m) near drainage canals, which are key fire ignition risk locations in the peatlands. Our study also shows that, on average, elevation and precipitation are the most important parameters influencing groundwater level spatiotemporally.

Estimating mangrove forest gross primary production by quantifying environmental stressors in the coastal area

Mangrove ecosystems play an important role in global carbon budget, however, the quantitative relationships between environmental drivers and productivity in these forests remain poorly understood. This study presented a remote sensing (RS)-based productivity model to estimate the light use efficiency (LUE) and gross primary production (GPP) of mangrove forests in China. Firstly, LUE model considered the effects of tidal inundation and therefore involved sea surface temperature (SST) and salinity as environmental scalars. Secondly, the downscaling effect of photosynthetic active radiation (PAR) on the mangrove LUE was quantified according to different PAR values. Thirdly, the maximum LUE varied with temperature and was therefore determined based on the response of daytime net ecosystem exchange and PAR at different temperatures. Lastly, GPP was estimated by combining the LUE model with the fraction of absorbed photosynthetically active radiation from Sentinel-2 images. The results showed that the LUE model developed for mangrove forests has higher overall accuracy (RMSE = 0.0051, R² = 0.64) than the terrestrial model (RMSE = 0.0220, R² = 0.24). The main environmental stressor for the photosynthesis of mangrove forests in China was PAR. The estimated GPP was, in general, in agreement with the in-situ measurement from the two carbon flux towers. Compared to the MODIS GPP product, the derived GPP had higher accuracy, with RMSE improving from 39.09 to 19.05 g C/m²/8 days in 2012, and from 33.76 to 19.51 g C/m²/8 days in 2015. The spatiotemporal distributions of the mangrove GPP revealed that GPP was most strongly controlled by environmental conditions, especially temperature and PAR, as well as the distribution of mangroves. These results demonstrate the potential of the RS-based productivity model for scaling up GPP in mangrove forests, a key to explore the carbon cycle of mangrove ecosystems at national and global scales.

An Overview of Remote Sensing Data Applications in Peatland Research Based on Works from the Period 2010–2021

In the 21st century, remote sensing (RS) has become increasingly employed in many environmental studies. This paper constitutes an overview of works utilising RS methods in studies on peatlands and investigates publications from the period 2010–2021. Based on fifty-nine case studies from different climatic zones (from subarctic to subtropical), we can indicate an increase in the use of RS methods in peatland research during the last decade, which is likely a result of the greater availability of new remote sensing data sets (Sentinel 1 and 2; Landsat 8; SPOT 6 and 7) paired with the rapid development of open-source software (ESA SNAP; QGIS and SAGA GIS). In the studied works, satellite data analyses typically encompassed the following elements: land classification/identification of peatlands, changes in water conditions in peatlands, monitoring of peatland state, peatland vegetation mapping, Gross Primary Productivity (GPP), and the estimation of carbon resources in peatlands. The most frequently employed research methods, on the other hand, included: vegetation indices, soil moisture indices, water indices, supervised classification and machine learning. Remote sensing data combined with field research is deemed helpful for peatland monitoring and multi-proxy studies, and they may offer new perspectives on research at a regional level.