The Impact of Agricultural Land Conversion on Carbon Stocks Across Canada, With a Focus on Grazing Lands

Background

The grazing lands and soil resources that support Canadian beef production play an important role in regional and global climate regulation via soil carbon (C) storage. These lands are finite and at risk of further reduction or loss due to irreversible conversion to other agricultural and non-agricultural land uses. The loss and degradation of grazing lands can result in negative consequences for soil C stocks and the ability of these soils to mitigate GHG emissions from multiple sectors of the economy, not just agriculture.

Given the potentially significant role of grazing lands across Canada in providing climate regulation services and mitigating GHG emissions from agricultural production through soil C storage under sustainable grazing practices, a comprehensive assessment of the consequences of agricultural land management, including conversion to other land uses, on soil C dynamics across space and time was crucial.

Objectives 

Overall objective: to quantify the impact, over multiple decades, on soil C dynamics of land cover change and beef management practices across Canada’s agricultural landscape:

• Assessment of land cover change at the ecodistrict-scale across Canada’s agricultural zone for three time periods: c. 1995-2011, 2011-2016 and c. 1995-2016, to produce spatially-explicit estimates of past and present soil C stocks on Canadian agricultural lands;
• Development of a soil organic carbon (SOC) database for agricultural lands in Canada using field-based measurements and development of a predictive SOC model to facilitate the estimation of SOC stocks for areas where no field measurements exist;
• Examination of the potential consequences of this land cover change for the soil C stock of agricultural lands from c. 1995-2016;
• Characterization of “typical” or “model” beef farms for different Canadian regions, using existing data and expert knowledge – each farm is representative of different environmental and management scenarios across Canada’s beef-producing area;
• Modelling of the long-term impact of each “model” farm on soil C dynamics and exploration of the potential impacts on soil C of specific alternative management practices.

What they DID

This project quantified changes in the areal extent of different agricultural land cover types – annual crops, (tame) pasture and forages and (native) grassland – throughout Canada’s agricultural zone and across three time points (c. 1995, 2011 and 2016) using multiple satellite-derived land cover maps. These maps were then coupled with field estimates of soil organic carbon stocks, compiled using an extensive and exhaustive literature search, to quantify the amount of carbon held in the agricultural soils of these different land cover types at each point in time, and the potential changes in these stocks due to shifts in land cover between time points. To facilitate the estimation of SOC stocks for areas where no field SOC measurements were available from the literature, the SOC database was also used to develop a predictive SOC model for the entire agricultural zone, based on SOC predictors including precipitation, temperature, soil characteristics, tillage regime and land cover type.

To examine soil C dynamics in greater detail and in response to a wider range of beef management practices, several “model” beef farms were characterized, based on existing data and expert knowledge, with each farm broadly representative of a different Canadian beef-producing region and its characteristic environmental and management conditions (e.g., Brown Soil Zone, Lake Manitoba Plain, etc.). Some of these farms have already been created in the Holos V4 interface. Holos V4 is a tool for estimating greenhouse gas emissions from Canadian crop and livestock operations, and also contains the Introductory Carbon Balance Model (ICBM), which simulates annual changes in the SOC stocks of the farm’s lands in response to management practices.

What they Learned

Significant shifts in the area of different agricultural land cover categories took place across Canada over the 20-year study period. Some of the most drastic gains/losses took place in areas important for the beef industry, including Alberta, where an estimated decline of 4.62 M ha of grazing lands was reported, mainly due to large losses of (tame) pasture & forage land, and Saskatchewan where a modest increase in grazing land (of ~19,500 ha) was reported, which was dwarfed by the fall in annual cropland area of just over 1.5 M ha.

An exhaustive literature search for per area SOC stock estimates for annual cropland, pasture & forages and (native) grassland across Canada yielded a total of 2,418 individual SOC observations that were included in the final SOC database. Overall, this exercise highlighted a distinct lack of SOC stock data for BC and Atlantic Canada, with the highest concentration of study sites and SOC measurements in the Prairie region, and in particular in Alberta.

Coupling these SOC estimates with land cover data for c. 1995 and 2016 gave some indication of potential shifts in agricultural SOC stocks that may have occurred following changes in the area of annual cropland, pasture & forage and grassland across 15 Canadian ecoregions. Although uncertainties in these estimates exist, some tentative conclusions can be drawn. For instance, across the seven Alberta ecoregions examined, widespread declines in the area of grazing lands (pasture & forage + grassland) c. 1995-2016 resulted in overall losses in SOC stocks, as it appears that at least some of the lost grazing lands were replaced by an expansion in annual cropland area, which was generally associated with lower per area SOC values. In Saskatchewan c. 1995-2016 period, for all four ecoregions examined net losses of total agricultural SOC were estimated due to declines in the overall agricultural area.

What it Means

Over a 20-year period, fluctuations in the area of different agricultural land cover types vital for the Canadian beef sector have occurred, with potentially significant consequences for the stocks of C contained within the soils of these lands. Although uncertainties exist in the analysis due to multiple factors, including uncertainties relating to the identification and classification of different agricultural land cover types (particularly perennial vegetation types categories) both within and between datasets and the paucity of field SOC measurements for certain regions and land cover classes, we can nonetheless infer that declines in agricultural land area and particularly in the area of perennial vegetation cover has had a negative impact on the agricultural SOC resource in some regions. The development of the predictive SOC model should allow us to overcome some of the issues associated with insufficient literature-derived SOC values, and to estimate SOC stocks for a larger portion of the agricultural land base.