Grasslands

Key Finding 2

Increasing temperatures and rising atmospheric carbon dioxide (CO₂) concentrations interact to increase productivity in northern North American grasslands, but this productivity response will be mediated by variable precipitation, soil moisture, and nutrient availability (high confidence, very likely).

Description of evidence base

Experimental manipulations in the field provide evidence of climate change effects on grassland productivity by up to 33%, but this is contingent on nutrient and moisture availability (e.g., Morgan et al., 2011; Mueller et al., 2016; Reich and Hobbie 2013). Spatially distributed observations of vegetation phenology (i.e., greenness) and carbon fluxes combined with empirical modeling provide evidence of regional differences in grassland responses to future climate change (Hufkens et al., 2016). Simulation models are in general agreement with empirical evidence that carbon stocks will increase in grasslands in the coming three to four decades (Zhu et al., 2011). In grasslands, carbon uptake is responsive to precipitation amount and event size and timing, with both positive and negative effects documented, but droughts are associated with carbon losses across all grasslands (Cherwin and Knapp 2012; Goldstein and Suding 2014; Heisler-White et al., 2008, 2009; Knapp et al., 2008b; Kulmatiski and Beard 2013; Thomey et al., 2011).

Major uncertainties

The largest source of uncertainty is related to future precipitation regimes in the grassland biomes of North America, with both increases and decreases in precipitation predicted (IPCC 2013). The degree to which altered precipitation regimes will affect carbon cycling in future grasslands is uncertain (Knapp et al., 2008b). The relative response of grassland productivity to moisture availability is contingent upon prior conditions, which vary temporally and spatially (Heisler-White et al., 2009). Empirical models represent grassland phenology and productivity well, but they lack explicit physiological processes, leading to uncertainties in mechanisms underlying ecosystem responses to climate change (Hufkens et al., 2016).

Confidence is high that grassland production will increase with precipitation as atmospheric CO₂ and temperature increase in the coming three to four decades, based on empirical evidence from field experiments.

If grassland productivity decreases in response to climate change, such as reduced precipitation, forage production for livestock is very likely to be at risk. This has been demonstrated by numerous experiments and models as explained above in the description of evidence base.

Summary sentence or paragraph that integrates the above information

Grassland productivity is highly likely to respond positively to increased precipitation and temperature, especially in the Northern Great Plains. Neutral or negative responses of productivity to warming in the Southern Great Plains, the southwestern United States, and Mexico may be offset by positive responses to elevated CO₂.

Key Finding 3

Soil carbon in grasslands is likely to be moderately responsive to changes in climate over the next several decades. Field experiments in grasslands suggest that altered precipitation can increase soil carbon, while warming and elevated CO₂ may have only minimal effects despite altered productivity (medium confidence, likely).

Description of evidence base

Meta-analysis of numerous field experiments showed that soil carbon stocks increase when precipitation is increased or decreased in grasslands (Zhou et al., 2016). Meta-analysis also showed that elevated CO₂ increased soil carbon decomposition rate, limiting carbon storage potential (van Groenigen et al., 2014). Field experiments indicate that soil carbon stocks decrease with warming, especially in regions where stocks are high to begin with (Crowther et al., 2016), although warming-induced soil carbon losses from grasslands may be insignificant (Lu et al., 2013). These results are confirmed in some simulation experiments (e.g., Parton et al., 2007; Shi et al., 2015).

Major uncertainties

Major uncertainties in soil carbon storage come from insufficient understanding of physical and biological mechanisms that determine the stability of soil carbon. Physical mechanisms underlying carbon stability in soil, such as protection within aggregates and their sensitivity to climate change, are still poorly described (Heimann and Reichstein 2008). In particular, regulation of soil organic matter decomposition by microbe-plant interactions is poorly understood and not well represented in models (Wieder et al., 2015). Improving mechanistic understanding of soil carbon dynamics, and incorporating key mechanisms into models, will reduce uncertainties in future carbon cycle predictions (Todd-Brown et al., 2013).

Assessment of confidence based on evidence and agreement, including short description of nature of evidence and level of agreement

Mechanistic understanding of soil carbon stability in the face of climate change is still limited, leading to only medium confidence levels regarding the response of soil carbon to climate changes.

Estimated likelihood of impact or consequence, including short description of basis of estimate

Soils in grasslands are not likely to respond strongly to climate change; small carbon losses or gains could occur in the future with warming or elevated CO₂. Larger carbon gains are likely to occur with increased precipitation.

Summary sentence or paragraph that integrates the above information

Mechanisms regulating soil carbon storage in response to climate change can be incorporated into models to improve confidence in model predictions of future carbon cycling.

Key Finding 4

Carbon stocks and net carbon uptake in grasslands can be maintained with appropriate land management including moderate levels of grazing. Fire suppression can lead to encroachment of woody vegetation and increasing carbon storage in mesic regions, at the expense of grassland vegetation (high confidence, likely).

Description of evidence base

Studies of carbon fluxes using eddy covariance indicate that moderate grazing allows grasslands to continue to be net carbon sinks, but heavy grazing diminishes their capacity to take up carbon (Frank 2004; Morgan et al., 2016; Polley et al., 2008; Risch and Frank 2006). Soil inventory studies indicate that moderate to light grazing does not negatively affect carbon stocks (Conant et al., 2001, 2017), and improving grazing management can augment carbon stocks (Chambers et al., 2016). Carbon cycle responses to woody encroachment are determined from inventories of carbon stocks in vegetation and soils in plots that have been experiencing woody encroachment for different periods of time (Barger et al., 2011; Knapp et al., 2008a).

Major uncertainties

Uncertainties in grazing management impacts on carbon cycling in grasslands stem mainly from the regional variations in soil carbon responses to management, from challenges in designing scientific studies that adequately represent real-world management practices, and from limitations faced when extrapolating plot-level studies to broader areas (Conant et al., 2017). Interactive effects of grazing, climate, soil type and plant community composition on carbon storage are not well constrained (McSherry and Ritchie 2013). The magnitude of carbon accumulation below ground in response to woody encroachment is poorly constrained, but change in carbon pools above ground is well known (Barger et al., 2011; Knapp et al., 2008a). Fire regimes are changing with increasing temperatures and altered vegetation; uncertainties in future fire risk add uncertainty to projections of carbon budgets.

Assessment of confidence based on evidence and agreement, including short description of nature of evidence and level of agreement

There is high confidence with general agreement across several studies that moderate to light grazing will not have a negative impact on carbon cycling.

Estimated likelihood of impact or consequence, including short description of basis of estimate

Woody encroachment likely will lead to increased carbon storage in mesic grasslands.

Summary sentence or paragraph that integrates the above information

Carbon likely will continue to accumulate for the next several decades in grasslands if they are appropriately managed.

Supporting Evidence for Tables

Table 10.1 is based on Hayes et al. (2012). The areas for grasslands by countries and the continent are from the models and inventory analyses used in their study (see Table S10 in Hayes et al., 2012). The area for “Others” is smaller for the models than the inventory analysis mainly because the latter includes urban areas. Inventory estimates are the sum of livestock methane (CH₄) emissions + livestock carbon dioxide (CO₂) emissions + grassland net ecosystem exchange (NEE) for Canada and the United States. Taiga was excluded from Canada grassland NEE and livestock emissions. For Mexico, the number for “Others” was used because extracting grassland NEE was not possible. Atmospheric inversion models (AIMs) and land-surface models (LSMs) are from Table 2 in Hayes et al. (2012) and do not include CH₄ emissions or human settlement emissions. Thus, the AIM values of NEE for “Others” should be representative of grassland and pastureland NEE. Area estimate for grasslands: www.statista.com/statistics/201761/projection-for-total-us-grassland-area-from-2010.

Table 10.2 Carbon fluxes and stocks for grasslands and shrublands in the conterminous United States summarized from the LandCarbon project (landcarbon.org/categories). Values are averages of the A1B, A2, and B1 climate scenarios and estimated using the FOREcasting SCEnarios of land-use change (FORE-SCE) model and the Erosion-Deposition-Carbon-Model (EDCM), CENTURY, and PBN carbon models (Liu et al., 2012b, 2014; Zhu et al., 2011). Climate projections based on emissions scenarios used by the LandCarbon Project are considered to be similar to representative concentration pathway (RCP) scenarios (Knutti and Sedláček 2013). Negative fluxes indicate carbon losses from the ecosystem; positive fluxes indicate carbon gains by the ecosystem. The total flux is considered to be the net ecosystem carbon balance (NECB). Land-cover classification could be a source of differences. Flux towers mostly measure actual grassland and rangeland, whereas the General Ensemble Biogeochemical Modeling System (GEMS) includes both grassland and shrubland. The conterminous United States has about 1 million km² of grassland and 1.3 million km² of shrubland (from Liu et al. land-cover data). The area difference is notable. Land conversion to and from agriculture and permanent grassland loss to urban land all contribute to the total carbon number.