Page Banner

United States Department of Agriculture

Agricultural Research Service

What We Study
headline bar

The goal of the PHACE experiment is to understand how elevated CO2 and warming influence plants and soils.  Specifically, we hope to learn how these changes influence the following key components of mixed-grass prairie:

Water controls most processes in the semiarid mixed-grass prairie. Elevated CO2 can decrease transpiration and increase water use efficiency and water availability. Warming often increases both evaporation and transpiration.

Nitrogen often limits plant productivity and quality, and can become increasingly limiting with elevated CO2.

Plant productivity and quality determine the ability of mixed-grass prairie to support grazing by livestock and wildlife, which are in turn key to the sustainability of agriculture and biological diversity.

Carbon cycling determines whether mixed-grass prairie sequesters or releases CO2, and therefore whether it contributes to or mitigates climate change. As the substrate for heterotrophic microbes, carbon also affects mineralization and immobilization of other nutrients.

Plant communities both respond to and influence biogeochemical cycling. In mixed-grass prairie, warm-season and cool-season grasses comprise the majority of plant biomass, while forbs comprise most of plant diversity. Legumes and sub-shrubs are less common, but can respond strongly to elevated CO2.

Invasive plants reduce forage production and biological diversity, and can thrive following increases in resource availability, including CO2 and water. PHACE includes several sub-experiments designed to test relative responses of native and invasive plants.

Microbes mediate biogeochemical cycling, and the structure of microbial communities can control both carbon cycling and nutrient availability to plants.



1. In this semi-arid grassland, elevated CO2 effects on soil water content counteract the dessicating effect of warmer temperatures. As anticipated, warming induced soil drying, but elevated CO2 increased soil water content more than expected, by increasing plant water use efficiency. Surprisingly, the combination of elevated CO2 and elevated temperature resulted in no difference in soil water content between these “future” plots and control plots under present-day CO2 and temperature levels (Figure 1).

Figure 1. Responses of soil water content (SWC) to CO2 & warming. Average and s.e.m. (error bars) of volumetric SWC (5–25 cm depth) for plots exposed to present-day ambient CO2 and temperature (ct), 1.5/3 C day/night warming (cT),600 p.p.m.v. CO2 (Ct), and 600 p.p.m.v. CO2 and 1.5/3 C day/night warming (CT) (five replications per treatment). Data are averaged over days of year (DOY) 100–200, the early- to mid-growing season when soil water most limits productivity. Precipitation amounts for this same period are also presented. Significance (P ≤ 0.05) for main effects and year are given in the figure (Figure 1 from Morgan et al. 2011).

These soil water effects influenced plant biomass responses. Higher CO2 increased plant biomass, especially in dry years when water savings were most important to growth (2006-2008). Warming alone did not significantly affect total aboveground biomass, likely because potential increases in plant growth were limited by lower soil water. There were no net decreases in plant growth between present-day and future conditions (elevated CO2 and temperature).

The effect of higher CO2 on soil water appears to be the dominant driver of plant biomass responses in semi-arid grasslands. Relative to plots at ambient CO2, the increase in aboveground biomass at elevated CO2 is stronger at a more negative early-season soil matric potential (i.e., drier soil), both at PHACE and an earlier CO2 enrichment experiment (Morgan et al. 2007). This relationship from two different ecosystems suggests that CO2 will increase plant productivity most when plants are actively growing but water-limited (Figure 2).

Figure 2. Response of biomass enhancement ratio to soil matric potential. Effects of early-season (DOY 100–200) rooting zone soil matric potential (ψm) on biomass enhancement ratio, the ratio of mid-July harvested above-ground plant biomass in CO2-enriched plots divided by plant biomass from ambient CO2 plots. Data are from the PHACE experiment (4 years’ data, n = 4), and from a previous open top chamber CO2 enrichment experiment (5 years’ data, n = 5) conducted on Colorado shortgrass steppe. For further details, see Supplementary Appendix I, soil water conversions (Figure 3 from Morgan et al. 2011).

For more results, see: Morgan, J.A. et al. 2011. C4 grasses prosper as carbon dioxide eliminates dessication in warmed semi-arid grassland. Nature 476: 202-206. and Morgan, J. A., et al. 2007. Carbon dioxide enrichment alters plant community structure and accelerates shrub growth in the shortgrass steppe. Proc. Natl Acad. Sci. USA 104, 14724–14729.

2. Elevated CO2 and warming have contrasting effects on N availability as well as moisture and productivity (Figure 3). Under elevated CO2, soil inorganic N decreased, likely due to increased microbial N immobilization. The CO2-induced increase in soil moisture facilitated higher N uptake by microbes, but did not affect plant N pool sizes.  In contrast, warming increased soil inorganic N and plant N.  Direct effects of warming on net N mineralization appeared to be more important than a warming-induced decrease in soil moisture. 

Figure 3. Soil inorganic nitrogen (N) availability (NH4+ + NO3−) at 2–7.6 cm soil depth in 2006 averaged by CO2, and in 2007 and 2008 averaged by CO2, warming, and irrigation treatments. Treatments: ct, ambient CO2 and ambient temperature; cT, ambient CO2 and elevated temperature; Ct, elevated CO2 and ambient temperature; CT, elevated CO2 and elevated temperature; ct-i, ambient CO2 and ambient temperature, but irrigated. Soil inorganic N availability is expressed in μg N 10 cm−2 resin membrane area of the plant root simulator (PRS) probes per month incubation time. Error bars indicate ± 1 SE. ANOVA P-values are reported when P < 0.05 (in bold) or P < 0.1 (in italics) (Figure 2 from Dijkstra et al. 2010).

For more results, see: Dijkstra, F.A. et al. 2010. Contrasting effects of elevated CO2 and warming on nitrogen cycling in a semiarid grassland. New Phytologist 187: 426-437. and Carillo et al. 2012. Controls over soil nitrogen pools in a semiarid grassland under elevated CO2 and warming.  Ecosystems 15: 761–774.

3. Wetter soil conditions under elevated CO2 increased P availability to plants and microbes relative to that of N, while drier conditions with warming reduced P availability relative to N (Figure 4). Soil moisture exerts an important control on inorganic P supply from desorption and dissolution reactions. This means that despite the fact that warming may alleviate N limitation under elevated CO2, warming and drought can exacerbate P limitation on growth and microbial activity.

Figure 4. N:P ratios in green (a, e) and senesced (b) plant tissue, microbes (c) and on Plant Root Simulator resin probes (PRS probes) (d) in response to elevated CO2, warming and irrigation (ct, ambient CO2 and temperature; cT, ambient CO2 and 1.5/3 C day/night warming; Ct, 600 ppmv CO2 and ambient temperature; CT, 600 ppmv CO2 and 1.5/3 C day/night warming; ct-i, ambient CO2 and temperature with 60 mm yr-1 irrigation). Green plant N:P ratios are species-weighted averages. Error bars represent 1SE. ANOVA P values are reported when P < 0.05 (in bold) or P < 0.1 (in italics) (Figure 1 from Dijkstra et al. 2012).

For more results, see: Dijkstra, F.A. et al. 2012. Climate change alters stoichiometry of phosphorus and nitrogen in a semiarid grassland. New Phytologist 196:807-815.

4. The PHACE plots were warmed by a free-air, feedback controlled infra-red heater system with target warming of 1.5/3.0 C day/night. We assessed how well the system maintained these target temperatures. Thermocouple temperature sensors were placed on leaves, in the soil, and above and within the plant canopy to monitor the actual warming at ecologically important zones.

Figure 5. The infra-red heater warming system worked very well. The leaf and soil temperatures were warmed very close to the day/night target warming of 1.5/3.0 C (note the green differential temperature bars). The warming system also performed well under very windy Wyoming conditions, and during both summer and winter.

For more results on the performance of the infra-red warming system see LeCain et al. 2015. Microclimate Performance of a Free-Air Warming and CO2 Enrichment Experiment in Windy Wyoming, USA. PLoS ONE. 10(2): e0116834. doi:10.1371/journal.pone.0116834


The data from the manuscript are available here: TFACE Solstice Power Use, Temperature Data, Solstice temperatures, Wind speed vs. Temp differences.

Last Modified: 4/6/2015
Footer Content Back to Top of Page