Kevin G. Harrison, Ph. D. Research: opening new vistas Finding the missing carbon I research processes that influence past, present and future atmospheric carbon dioxide levels. I have developed a strategy for quantifying the amount of additional carbon stored in soil due to carbon dioxide fertilization. The initial results from this soil carbon research help explain why atmospheric carbon dioxide levels are increasing more slowly than expected. I have pioneered the "silica hypothesis" that suggests that increasing the inventory of silica in the ocean may increase diatom abundance while decreasing coccolith abundance. Recently, I have introduced the “reverse ocean acidification hypothesis” that suggests that the soft-tissue pump efficiency increased when surface ocean pCO2 levels decreased. Both these changes help explain why atmospheric carbon dioxide levels were lower during glacial times. Soil carbon research Active soil carbon My first research objective has been to quantify the amount of carbon that exchanges between soil and the atmosphere. Without this information, it is impossible to determine if carbon stored in soil could significantly change carbon dioxide levels and if soil could be the location of the “missing sink.” The “missing sink” is the term coined by scientists to describe the contemporary imbalance between the known sources and sinks of atmospheric carbon dioxide. About 25% of the carbon dioxide released to the atmosphere by fossil fuel combustion and changing land use is missing. This “missing sink” is slowing the build-up of carbon dioxide in the atmosphere. My soil radiocarbon research has shown that soil carbon exchanges between 20 and 25 billion tons of carbon/year with the atmosphere (Harrison et al., 1993a; Harrison, 1996; Harrison and Bonani, 2000). Just under half of the terrestrial net primary production resides in the soil before being returned to the atmosphere. As the first quantitative estimate of this flux, my research has clearly demonstrated soil carbon’s potential to influence atmospheric carbon dioxide levels and be the possible location of the “missing sink.” Soil contains carbon that has turnover times ranging from days to thousands of years. To understand these dynamics, my research team developed an analytical method to isolate mineral-bound soil carbon (Mahoney et al., 2003). This enabled bulk radiocarbon measurements to be used to determine the inventory and turnover time of both active soil carbon and passive soil carbon. These turnover times and inventories are fully constrained by carbon and radiocarbon measurements. It was surprising to discover that soil carbon compounds having a multitude of turnover times could be represented by a twocomponent model and that switching to a three-component model resulted in poorer agreement between the model and the data. Soil consists of a mixture of active and passive carbon, but it is only the active carbon that responds to perturbations. Although experiments that measure changes in bulk carbon in response to perturbations are interesting, their value and usefulness increase exponentially if they determine the initial and final active soil carbon turnover times and inventories as I illustrate below. A second benefit of isolating the mineral-bound soil was that the

Harrison

1

variability in the mineral-bound soil carbon was less than the variability observed in the bulk soil carbon. Dynamic carbon storage My research has introduced the concept of “dynamic carbon storage” (Harrison, 1993a; Harrison, 2004; Harrison et al., 2004). Dynamic carbon storage occurs in pools having short turnover times of 100 years or less. The carbon accumulates in the active carbon pool because the flux of the carbon into the pool exceeds the flux of the carbon leaving the pool. Excepting fossil fuels, my research shows that carbon pools having turnover times of greater than 1000 years cannot significantly change atmospheric carbon dioxide levels in our lifetime (Harrison et al., 1993a). My colleagues and I have tested my theories by seeing how soil carbon dynamics and inventories respond to perturbations. We discovered that cultivated soil had lower radiocarbon values than native soil, because plowing mixes radiocarbon-rich surface soil with depleted deeper soil, and farming reduces the inventory of active soil carbon (Harrison et al., 1993b). A related study explained changes in soil carbon and radiocarbon following agricultural abandonment. We observed a 12-year turnover time for carbon in recovering soil, which is twice as fast as the soil carbon turnover time in undisturbed temperate forests and grasslands (Harrison et al., 1995). The observed rapid turnover demonstrates that soil can remove carbon dioxide from the atmosphere faster than expected. It also shows that recovering soil will respond to perturbations, such as elevated carbon dioxide levels in the atmosphere, faster than native soil. Carbon turnover times can be used to estimate the flux of atmospheric carbon dioxide into abandoned agricultural land, a number that was not well known before this approach was developed. My approach has demonstrated that active soil carbon has the potential to influence atmospheric carbon dioxide levels and is one of the most likely locations for the “missing sink.” Processes that could alter the amount of active soil carbon storage include CO2 fertilization, changing land use, anthropogenic nitrogen deposition, and climate change. CO2 fertilization My next objective was to see if CO2 fertilization has been removing enough carbon dioxide from the atmosphere and storing it in the active soil carbon reservoir to balance the global budget. CO2 fertilization occurs if vegetation grows faster when exposed to higher carbon dioxide levels. The accelerated plant growth may increase belowground soil carbon storage. To test the CO2 fertilization/active soil carbon storage hypothesis, I looked at soil carbon changes in three enrichment experiments. I have found that elevated carbon dioxide levels increased soil carbon storage in three experimental settings—an intact forest Free-Air Carbon Enrichment (FACE) site, a white oak experiment, and a loblolly experiment—where trees were exposed to elevated CO2 levels. My research team found that soil carbon accumulation rates for an intact forest exposed to 570 ppm of carbon dioxide were 30% greater than the forest’s ambient counterpart after 3.5 years (Harrison et al., 2001; Heumann et al., 2001); that soil carbon accumulation rates for white oaks exposed to 660 ppm of carbon dioxide were 14% greater than their ambient counterparts after four years (Harrison et al., 2004); and that soil carbon accumulation rates for loblolly pines exposed to 660 ppm of carbon dioxide were 21% greater than their ambient

Harrison

2

counterparts after four years (Grunauer et al., 1998). The results of these three studies have been supported by similar results from CO2 fertilization studies reported by other researchers. For example, Luo et al. (2006) looked at 104 published papers from various carbon dioxide fertilization experiments. They found forty experiments that reported soil carbon results. The soil collected from the elevated sites had, on average, 5.4% more carbon than soil collected from the ambient sites. CO2 fertilization factor Although these empirical results are interesting, they do little to improve predictions of future atmospheric carbon dioxide levels or global warming. To add value to these findings, I have developed the concept of the soil carbon CO2 fertilization factor (σ CF) (Harrison, 2004). The σ CF lets researchers compare the results of carbon dioxide enrichment experiments that have different soil carbon turnover times, different levels of CO2 enrichment, and different lengths of exposure to elevated carbon dioxide levels. I have used the σ CF to estimate increases in soil carbon uptake due to observed contemporary increases in atmospheric carbon dioxide levels. I calculated σ CF for each experimental setting by measuring changes in carbon inventories and radiocarbon ratios and determining the inventories and turnover time of active soil carbon. The intact forest had a σ CF of 1.8. The white oak ecosystem had a σ CF of 1.18 (Harrison et al., 2004; Harrison, 2004) and the loblolly pine stand had a σ CF of 0.9. If the average soil carbon CO2 fertilization factor was 0.7, it would account for the “missing sink.” These results show that elevated carbon dioxide levels in the atmosphere are increasing the flux of carbon from the atmosphere to soil. Concurrent with my team’s CO2 fertilization research, my colleagues and I developed the experimental technique and theoretical background to differentiate between carbon dioxide respired by roots and carbon dioxide respired by microbial oxidation (Andrews et al., 1999). My colleagues and I found that root respiration comprised 55% of the total soil respiration. Future soil carbon research My future research will include measuring the CO2 fertilization factor for major ecosystems across the globe and determining how climate change alters active soil carbon storage. My approach can also be used to test and refine soil carbon fractionation methods and provide parameters for carbon models that include carbon fluxes to and from soil. My research has provided a new approach for examining how terrestrial ecosystems respond to perturbations, such as CO2 fertilization, climate change, anthropogenic nitrogen deposition, and changing land use. The increased flux of carbon into soil may increase nitrogen fixation and accelerate chemical weathering, so nutrient limitations may not significantly alter the observed CO2 fertilization response. For example, we observed a large increase in soil nitrogen below the intact forest that was exposed to elevated atmospheric carbon dioxide levels and will investigate how weathering rates change in response to elevated carbon dioxide levels. Nitrogen may also be supplied by anthropogenic nitrogen deposition in some areas. My research team has started to look at how elevated carbon dioxide levels influence carbon storage in litter. It will be interesting to see how changing soil carbon input affects populations of nitrogen fixing bacteria, nitrogen fixation rates, and the weathering rates of various elements. During glacial times, the lower carbon dioxide levels in the atmosphere may have caused “CO2 starvation.” My research team

Harrison

3

will also look at how “CO2 starvation” impacted terrestrial carbon storage and rock weathering rates. Silica research My “silica” and “reverse ocean acidification” hypotheses may help explain why carbon dioxide levels were lower during glacial times and why atmospheric radiocarbon levels were higher during glacial times (Harrison, 1995, 2000, 2010). Carbon dioxide levels were about 80 ppm lower during glacial times compared to interglacial times. The resulting improved understanding of the global carbon cycle will lead to more accurate predictions of future carbon dioxide levels. Silica hypothesis The “silica hypothesis” suggests that an increase in dust delivery to the ocean set into motion a series of processes that drastically changed the chemistry, biology, physics and pCO2 of the glacial ocean. The glacial ocean was very different from the modern ocean and the principle of uniformitarianism needs to be overlooked to understand why carbon dioxide levels were lower during glacial times. I have shown that dust levels would only need to have increased by a factor of 2 to 7 to increase the oceanic inventory of silica enough to cut calcite production by 40% (Harrison, 2000). Later research by my team has shown that the amount of silica released by dust is greater than estimates used in Harrison (2000), so the actual amount of dust required to support the decrease in calcite has decreased. It’s clear that dust could provide enough silica to the ocean to increase the inventory of silica in the deep ocean, which, in turn, increases the flux of silica into the mixed layer. In the presence of sufficient silica, diatoms out-compete coccoliths in the mixed layer, which reduces the production of calcite. The presence of iron in the mixed layer causes diatoms to use the silica more efficiently, also increasing diatom populations. The resulting thinner diatom shells increase silica recycling efficiency in the upper ocean and hinder the preservation of diatom shells in the marine sediments. Reverse Ocean Acidification hypothesis During interglacial times, the concentration of carbon dioxide in the atmosphere averaged 280 ppm. In contrast, carbon dioxide levels were about 90 ppm lower during the last glacial maximum, which enabled phytoplankton to grow larger and denser calcium carbonate shells. This concept shows how these larger and denser shells may have helped maintain low atmospheric carbon dioxide levels during glacial times by increasing the efficiency of the softtissue pump. Future silica research Future silica hypothesis research includes measuring sterols and alkenones in marine sediment to document the increase in diatoms during glacial times, doing additional dust dissolution

Harrison

4

experiments to see if Si isotopes are fractionated by the dissolution process, and experimentally testing many aspects of the “silica” and “reverse ocean acidification” hypotheses to further develop these ideas. Dissolved organic carbon research Currently, scientists do not know the turnover time of dissolved organic carbon in seawater. Without this knowledge, it is impossible to determine if the relatively large inventory of dissolved organic carbon has the potential to influence atmospheric carbon dioxide levels. I have developed a technique for using radiocarbon measurements to determine the turnover time of soil organic matter (Harrison et al., 1993a). I will modify and use this approach to determine the turnover time of dissolved organic carbon in seawater. Summary My research is curiosity-driven. I love learning and solving problems. My research has helped explain why contemporary carbon dioxide levels in the atmosphere are increasing more slowly than expected and why carbon dioxide levels were lower during glacial times. In the future, I hope to do more experimental research to extend my soil carbon and silica research. My research builds on the work of other researchers and would not be possible without the generous support of mentors and colleagues. I hope that my research will help other scientists extend their research.

Harrison

5

References Andrews, J.A., K.G. Harrison, R. Matamala & W.H. Schlesinger. 1999. Separation of root respiration from total soil respiration using C-13 labeling during Free-Air CO2 Enrichment (FACE). Soil Science Society of America Journal, 64, 1429-1435. Grunauer, M.A., K.G. Harrison, A.L. Kafka, D.T. Tissue & R.B. Thomas. 1998. Measuring the effect of CO2 fertilization on soil organic material beneath Loblolly pines. EOS Trans., American Geophysical Union, 79, 17, S47. Harrison, K. G. 1995. The role of increased silica input on Paleo-CO2 levels. EOS, 76, 46, F292. Harrison, K.G. 1996. Using bulk soil radiocarbon measurements to estimate soil carbon turnover times: Implications for atmospheric CO2 levels. Radiocarbon, 38, 3, 181-190. Harrison, K.G. 2000. The role of increased marine silica input on paleo-pCO2 levels. Paleoceanography, 15, 3, 292-298. [Reviewed by Treguer, P. and P. Pondaven. 2000. Silica control of carbon dioxide. Nature, 406, 358-359.] Harrison, K.G. 2004. The soil carbon CO2 fertilization factor: The measure of an ecosystem's capacity to increase soil carbon storage in response to elevated CO2 levels. Geochemistry, Geophysics, Geosystems, 5, 5, Q05002, doi:10.1029/2003GC000686. Harrison, K. G. 2010. Can changes in oceanic biogeochemical cycles explain atmospheric carbon dioxide levels and radiocarbon levels during the last glacial maximum?, Eos Trans., 91(26), Ocean Sci. Meet. Suppl., Abstract PO21C-02. Harrison, K.G. & G. Bonani. 2000. A strategy for estimating the potential soil carbon storage due to CO2 fertilization. In The Global Carbon Cycle, edited by T. M. L. Wigley, D. S. Schimel, Cambridge: Cambridge University Press, 141-150. Harrison, K.G., W.S. Broecker & G. Bonani. 1993a. A strategy for estimating the impact of CO2 fertilization on soil carbon storage. Global Biogeochemical Cycles, 7, 1, 69-80. Harrison, K.G., W.S. Broecker & G. Bonani. 1993b. The effect of changing land use on soil radiocarbon. Science, 262, 725-726. Harrison, K.G., R.J. Norby, W.M. Post & E.L. Chapp. 2004. Soil carbon accumulation in a white oak CO2 enrichment experiment via enhanced root production. Earth Interactions, 8, 1-15, DOI: 10.1175/10873562(2004)8<1:SCAIAW>2.0.CO;2. Harrison, K.G., W.M. Post & D.D. Richter. 1995. Soil carbon turnover in a recovering temperate forest. Global Biogeochemical Cycles, 9, 4, 449-454.

Harrison

6

Harrison, K.G., L. Weeden, R.J. Heumann & A.L. Kafka. 2001. Using a FACE experiment to measure the amount of carbon transferred from the atmosphere to soil because of CO2 fertilization. EOS Trans., 82, 20, S77. Heumann, R.J., A.L. Kafka, & K.G. Harrison. 2001. Using a FACE experiment to measure the amount of soil carbon and nitrogen accumulation due to elevated atmospheric carbon dioxide levels. EOS, Trans., 82, 20, S92. Luo, Y., D. Hui & D. Zhang. 2006. Elevated CO2 stimulates net accumulations of carbon and nitrogen in land ecosystems: a meta-analysis. Ecology, 87, 1, 53-63. Mahoney, R. J., A.L. Kafka & K.G. Harrison. 2003. Procedure for determining soil-bound organic carbon and nitrogen. In Changing Land Use and Terrestrial Carbon Storage, Global Discovery Press, 1-21.

Harrison

7