Early Biosphere 2 Research

Significant Research Outcomes

Growing food on Mars, Not a novel idea for Biosphere 2:

The early goals of the project were to design an enclosed facility that could be used by humans to live on other planets, and to demonstrate the inter-connectedness of humans and the environment (Allen 1991).  Construction of the main apparatus began in January 1987 and was concluded in September 1991.  The first mission (September 1991 to September 1993)  in which four men and four women lived inside the totally sealed but, energy rich environment of Biosphere 2 growing all their food and recycling all their air, water and wastes (figure 1 attached).  The experiment was an outstanding success in engineering terms but failed as a sustainable planetary ecosystem analog.

The tightly closed structure (leak rate of less than 10% per year) (Dempster 1999) was furnished with an extremely rich organic soil (Leigh et al 1999, Scott 1999).  The soil supported rapid growth of the synthetic model ecosystems and crops in Biosphere 2.  However, the rich soil was the major factor in causing the experiment to become unsustainable.  Soil metabolism was so active, and soil reserves of carbon were so great, that atmospheric composition changed rapidly.  Oxygen was absorbed from the air by soil microbes and these released huge amounts of CO2 from the soil back to the air that exceeded the photosynthetic capacity of plants to assimilate it and to regenerate O2.  Instead, the excess CO2 was absorbed by the fresh, unsealed concrete of the structure, and O2 levels declined rapidly (Severinghaus et al 1994); O2 make-up was  needed (Dempster 1999) if the 8 humans were to survive (figure 2 attached).


Allen, J. (1991) Biosphere 2: The Human Experiment. Penguin Books NY (ISBN 0 14 01 5392 6).

Dempster, W. (1999) Biosphere 2 engineering design. Ecological Engineering special issue vol. 13, Nos. 1-4.

Leigh, L.S., Burgess, T., Marino, B.D.V., Wei, Y.D. (1999) tropical rainforest biome of Biosphere 2: Structure, composition and results of the first 2 year of operation. Ecological Engineering special issue vol. 13, Nos. 1-4.

Severinghaus, J.P., Broecker, W.S., Dempster, W.F., MacCallum, T., and Wahlen, M., (1994) Oxygen loss in Biosphere 2, EOS Trans. Amer. Geophys. Union 75, 33-37.

Scott, H. (1999) Characteristics of soils in the tropical rainforest biome of Biosphere 2 after 3 years. Ecological Engineering special issue vol. 13, Nos. 1-4

CO2 and Oceans

CO2 levels near 400ppm for the first time—Consequences for our ocean:

The Biosphere 2 marine mesocosm is a totally enclosed biome and was designed to simulates a Caribbean reef.  It is a large tank with a surface area of 35x20 m, from 7 m deep grading to a shallow lagoon partially separated by a fringing reef.  The total water volume of this mesocosm is 2.5x106 liters.  Mechanical systems simulate or substitute for natural environmental processes.  Physical and chemical parameters such as mixing, gas exchange, nutrient concentrations, and partial pressure of CO2 can be independently manipulated.   The Biosphere 2 ocean is ideal for testing models of chemical or biological changes on coral reefs.  The core experiment conducted in this facility was to measure the rates of photosynthesis and calcification for three states of the system.  One corresponding to full glacial time (CO2 = 200µatm, CO3= = 355 µmol/kg, pH 8.29), a second to the present day conditions (CO2 = 360 µatm, CO3= = 241 µmol/kg, pH = 8.08) and a third to conditions predicted for the 22nd Century (CO2 = 900µatm, CO3= = 174 µeq/kg pH 7.89)(Figure 1).  These conditions were alternated at 3-month intervals over a three year period. Results from Langdon et al (2000) have shown that rates of coral skeleton calcification decline by 40 % under seawater carbonate concentrations in equilibrium with CO2 in the mid 21st Century atmosphere.


Langdon C, Takahashi T, Marubini F, Atkinson MJ, Sweeney C, Aceves H, Barnet H, Chipman D, Goddard J (2000).  Effect of calcium carbonate saturation state on the calcification rate of an experimental coral reef.  Global Biogeochemical Cycles 14: 639-654

Langdon, C, Broecker W, Hammond D, Glenn E, Fitzsimmons K, Nelson SG, Peng TH, Hajdas I, Bonani G.  2003. "Effect of elevated CO2 on the community metabolism of and experimental coral reef." Global Biogeochemical Cycles 17(1): 11-1 to 11-14

El Niño Impacts

Biggest El Niño since 1979—Consequences for tropical rain forests: 

The rainforest mesocosm at the north end of B2L is a 35,000 m3 tropical rainforest habitat. (Leigh et al. 1999). The rainforest mesocosm has been growing in Biosphere 2 since 1989.  The rainforest biome presents a complex patchwork of differing species and different canopy structures that is ideal for evaluation of component photosynthetic activities  It is not practical to subdivide the rainforest at Biosphere 2 to permit a conventional experimental design with separate treatment and control plots.  Instead, time series experiments somewhat analogous to those in the ocean mesocosm were conducted.  The chamber was maintained at constant conditions for intervals of time while the rates of net CO2 exchange, transpiration, isotope exchange and trace gas exchanges were monitored. The magnitude and dynamics of the responses were followed. Different processes of the ecosystem will respond to changes in the environment with different relaxation times that reflect direct and indirect effects of the perturbation. One core experiment conducted in this mesocosm was to measure the net ecosystem exchange of CO2 in this system under different atmospheric CO2 concentration ranging from 380 to 820 ppm. Lin et al (1998, 1999, 2001) confirmed that the big left model predicted the Biopshere 2 rainforest response well . It suggests that tropical forest many be nearing their sink-capacity. 


Lin G, Marino BDV, Wei Y, Adams J, Tubiello F, Berry JA (1998). An experimental and model study of the responses in ecosystem exchanges to increasing CO2 concentrations using a tropical rainforest mesocosm. Australian Journal of Plant Physiology 25: 547-556

Lin G, Adams J, Farnsworth B, Wei Y, Marino BVD, Berry JA (1999).  Ecosystem carbon exchange in two terrestrial ecosystem mesocosms under changing atmospheric CO2 concentrations.  Oecologia 119: 97-108.

Lin G, Berry JA, Kaduk J, Griffin K, Southern A, Adams J, Van Haren J, Broecker W (2001) Sen­sitivity of photosynthesis and carbon sinks in world tropical rainforests to projected atmos­pheric CO2 and associated climate changes. Proceedings 121h International Congress on Photo­synthesis. CSIRO Publishing, Melbourne

CO2 and Isoprene Emissions

Increasing CO2 could mean plants give off more greenhouse gasses:  

The forestry biome coverd about 2,000 m2, has an air volume of 38,000 m3, and a soil volume of about 2,000 m3.  Soil depth averages about 1m. The forestry mesocosm was partitioned into 3 separate compartments each with its own atmospheric CO2 control system, environmental controls, sensors (temperature, relative humidity, light, and soil moisture) and continuous trace gas monitoring system. A forest of genetically identical cottonwoods (Populus deltoides Barr.) was planted in May 1998. These trees were coppiced yearly and allowed to grow from the stump. 

The impact of atmospheric CO2 on isoprene emissions.   Isoprene, an atmospherically reactive trace gas, is released by many plants, impacting the carbon balance of the planet and local atmospheric chemistry. When oxidized in the atmosphere above forested ecosystems in North America, isoprene contributes to the production of tropospheric ozone (a radiatively-active gas with important potential contributions to climate warming and a potentially destructive oxidative pollutant). Isoprene is synthesized from carbon compounds in the chloroplasts of plant leaves, and is thus potentially sensitive to increased photosynthesis rates caused by elevated CO2 concentrations. Thus, in a world of elevated CO2, the earth’s forest could emit more isoprene, causing the increased production of tropospheric ozone. Modeling ecosystem emissions of isoprene is an active area of research, and the three treatments of the Biosphere 2 experiment will be the first realistic test of how CO2 changes will affect ecosystem level emissions. To date, the connection between isoprene emission rate and elevated CO2 has been studied in potted plants, grown for short periods in elevated CO2 atmospheres. The results have been equivocal and difficult to interpret. Isoprene concentrations in air were monitored and converted to ecosystem level fluxes.  This top/down approach was combined with leaf and soil level measurements of isoprene fluxes to test current emissions models and gain mechanistic understanding of the drivers (CO2, leaf temperature and light) behind isoprene production.

The growth rate of these trees was dramatic, averaging a meter a month, and the plants were used to measure a number of responses to elevated CO2 (Griffin et al 2001, Griffin et al 2002, Rosentiel et al 2003, Engel et al 2004).

Griffin KL, Anderson OR, Gastrich MD, Lewis JD, Lin G-H, Schuster W, Seeman J, Tissue DT, Turnbull MH, Whitehead D (2001).  Plant growth in elevated CO2 alters mitochondrial number and chloroplast fine structure.  Proceedings of the National Academy of Sciences USA, 98: 2473-2478.

Griffin KL, Turnbull MH, Murthy R, Lin G-H, Adams J, Farnsworth B, Mahato T, Bazin G, Potosnak M, Berry JA (2002). Leaf respiration is differentially affected by leaf vs. stand-level night-time warming. Global Change Biology 8: 479-485

Griffin KL,Turnbull M, Murthy R (2002). The effect of canopy position on the temperature response of leaf respiration in Populus deltoides.  New Phytologist 154: 609-619

Rosenstiel, T.N., M.J. Potosnak, K.L. Griffin, R. Fall and R.K. Monson. 2003. Increased CO2 uncouples growth from isoprene emission in an agriforest ecosystem, Nature, 421:256-259.

Engel, V.C, K.L. Griffin, R. Murthy, L. Patterson, C.A. Klimas and M.J Potosnak. 2004. Growth CO2 modifies the transpiration response of Populus deltoides to drought and vapor pressure deficit, Tree Physiology, 24 (10): 1137–1145.