A significant contributor to CO2 emitted to the atmosphere comes from soil respiration (Rs).
Until recently, there have been no clear assessment of how much CO2 and the role that soils contribute in emissions relative to other greenhouse gases. Now a global scale map using statistical models and satellite imagery, along with other work by scientists, is beginning to indicate how much land use and soil change can affect our planet’s climate.
In a recent study, a series of machine learning models using multiple nonlinear regression (MNLR), random forest regression (RFR), support vector regression (SVR), and artificial neural network (ANN) were assessed for determining Rs datasets and their measures.
These models were used to create as accurate a picture as possible for soil respiration based on different global biomes evaluated, where the best results from these evaluations were used to make estimates. This allowed a global Rs dataset for 2000-2014 to be created. A global map measuring 1 x 1 km for each square was then created using composite satellite imagery. From that, over the period 2000-2014, maps of Rs are created for each square kilometer on our planet.
Overall, regions with high Rs are mainly in the tropics, including the Amazon Basin, Central Africa, and Southeast Asia, while low Rs are widely distributed in the Northern Hemisphere and mainly high-latitude areas in the western United States, Canada, Central Asia, parts of northern Mongolia, northeast China, Argentina, and Australia. Over time, Rs can be seen to increase in the northern latitudes at a more rapid pace, despite being generally lower, with increases approaching 8.5%.
From the results, it is clear that land-cover change was a major contributor to the changes in Rs in temperate (58%) and northerly boreal regions (55%).
Climate change itself is a driver of Rs change, as it affects plant respiration and plant growth. This is most pronounced at a general level at a global scale and in tropical regions, with impact of these factors ranging between 56% and 66% respectively.
It is evident that land use change was a greater factor for Rs change in boreal and temperate regions than in tropical regions.
Central United States, western Europe, northeast China, Kazakhstan, Argentina, east Brazil, east Africa, south Africa, and western Australia all showed decreasing Rs, which is attributed with drought conditions that have diminished respiration rates.
Overall, the results make it clear that land use is having a major change, particularly in the northern hemisphere, with how CO2 emissions are changing. Steps for conservation of different biomes can now better account for how much CO2 could be saved.
This type of work follows on recent advancements in the use of erosion and land use models that could be better coupled with global circulation climate models (GCMs) used to forecast temperature and precipitation change. Models such as the Revised Universal Soil Loss (RUSEL) and Soil Water Assessment Tool (SWAT) can now be better coupled with GCMs using spatially derived measures.
While studies have emphasized how much greenhouse emissions are occurring based on land use and its impact on climate, studies are also trying to demonstrate mitigation strategies that can improve soil conservation and diminish greenhouse emissions from improved preservation.
For instance, it was demonstrated that paludiculture, that is using wet agriculture and forestry in wetlands, can help keep Rs to lower levels. Such studies indicate that strategies created to lower Rs or different biomes may be needed and that these strategies will different for these biomes given the types of vegetation that can grow.
While many studies have applied spatial and computational methods to estimate greenhouse gas emissions, few studies have looked at global scale soil contribution and how changes to soil affect greenhouse emissions. Now that we have a picture of this, while models to estimate soil change have improved in integrating macro-scale climate with local change, we can also predict likely emission changes based on land use change for local regions.
This will be an important contribution and assistance in conservation efforts as strategies to limit greenhouse emissions gain global interest.
 For more on the detailed study looking at changing Rs at a global scale, see: Huang, N., Wang, Li, Song, X.-P., Black, T.A., Jassal, R.S., Myneni, R.B., Wu, C., Wang, Lei, Song, W., Ji, D., Yu, S., Niu, Z., 2020. Spatial and temporal variations in global soil respiration and their relationships with climate and land cover. Science Advances 6, eabb8508. https://doi.org/10.1126/sciadv.abb8508.
 For more on soil models and how they are improving in the use of GCMs coupled with them, see: Kumar, S., 2020. Geospatial Applications in Modeling Climate Change Impact on Soil Erosion, in: Venkatramanan, V., Shah, S., Prasad, R. (Eds.), Global Climate Change: Resilient and Smart Agriculture. Springer Singapore, Singapore, pp. 249–272. https://doi.org/10.1007/978-981-32-9856-9_12
 For more on strategies to diminish greenhouse emissions from soils in wetland regions, see: Tanneberger, F., Schröder, C., Hohlbein, M., Lenschow, † Uwe, Permien, T., Wichmann, S., Wichtmann, W., 2020. Climate Change Mitigation through Land Use on Rewetted Peatlands – Cross-Sectoral Spatial Planning for Paludiculture in Northeast Germany. Wetlands. https://doi.org/10.1007/s13157-020-01310-8.