The biogeochemistry of sulfur at Hubbard Brook

G. E. Likens, C. T. Driscoll, D. C. Buso, M. J. Mitchell, G. M. Lovett, S. W. Bailey, T. G. Siccama, W. A. Reiners, C. Alewell

Research output: Contribution to journalArticlepeer-review

179 Scopus citations


A synthesis of the biogeochemistry of S was done during 34 yr (1964-1965 to 1997-1998) in reference and human-manipulated forest ecosystems of the Hubbard Brook Experimental Forest (HBEF), NH. There have been significant declines in concentration (-0.44 μmol/liter-yr) and input (-5.44 mol/ha-yr) of SO42- in atmospheric bulk wet deposition, and in concentration (-0.64 μmol/liter-yr) an d output (-3.74 mol/ha-yr) of SO42- in stream water of the HBEF since 1964. These changes are strongly correlated with concurrent decreases in emissions of SO2 from the source area for the HBEF. The concentration and input of SO42- in bulk deposition ranged from a low of 13.1 μmol/liter (1983-1984) and 211 mol/ha-yr (1997-1998) to a high of 34.7 μmol/liter (1965-1966) and 479 mol/ha-yr (1967-1968), with a long-term mean of 23.9 μmol/liter and 336 mol/ha-yr during 1964-1965 to 1997-1998. Despite recent declines in concentrations, SO42- is the dominant anion in both bulk deposition and stream water at HBEF. Dry deposition is difficult to measure, especially in mountainous terrain, but was estimated at 21% of bulk deposition. Thus, average total atmospheric deposition was 491 and 323 mol/ha-yr during 1964-1969 and 1993-1998, respectively. Based on the long-term δ34S pattern associated with anthropogenic emissions, SO42- deposition at HBEF is influenced by numerous SO2 sources, but biogenic sources appear to be small. Annual throughfall plus stemflow in 1993-1994 was estimated at 346 mol SO42-/ha. Aboveground litterfall, for the watershed-ecosystem averaged about 180 mol S/ha-yr, with highest inputs (190 mol S/ha-yr) in the lower elevation, more deciduous forest zone. Weathering release was calculated at a maximum of 50 mol S/ha-yr. The concentration and output of SO42- in stream water ranged from a low of 42.3 μmol/liter (1996-1997) and 309 mol/ha-yr (1964-1965), to a high of 66.1 μmol/liter (1970-1971) and 849 mol/ha-yr (1973-1974), with a long-term mean of 55.5 μmol/liter and 496 mol/ha-yr during the 34 yrs of study. Gross outputs of SO42- in stream water consistently exceeded inputs in bulk deposition and were positively and significantly related to annual precipitation and streamflow. The relation between gross SO42- output and annual streamflow changed with time as atmospheric inputs declined. In contrast to the pattern for bulk deposition concentration, there was no seasonal pattern for stream SO42- concentration. Nevertheless, stream outputs of SO42- were highly seasonal, peaking during spring snowmelt, and producing a monthly cross-over pattern where net hydrologic flux (NHF) is positive during summer and negative during the remainder of the year. No significant elevational pattern in streamwater SO42- concentration was observed. Mean annual, volume-weighted soil water SO42- concentrations were relatively uniform by soil horizon and across landscape position. Based upon isotopic evidence, much of the SO42- entering HBEF in atmospheric deposition cycles through vegetation and microbial biomass before being released to the soil solution and stream water. Gaseous emissions of S from watershed-ecosystems at HBEF are unquantified, but estimated to be very small. Organic S (carbon bonded and ester sulfates) represents some 89% of the total S in soil at HBEF. Some 6% exists as phosphate extractable SO42- (PSO4). About 73% of the total S in the soil profile at HBEF occurs in the Bs2 horizon, and some 9% occurs in the forest floor. The residence time for S in the soil was calculated to be ∼9 yr, but only a small portion of the total organic soil pool turns over relatively quickly. The S content of above- and belowground biomass is about 2885 mol/ha, of which some 3-5% is in standing dead trees. Yellow birch, American beech and sugar maple accounted for 89% of the S in trees, with 31% in branches, 27% in roots and 25% in the lightwood of boles. The pool of S in living biomass increased from 1965 to 1982 due to biomass accretion, and remained relatively constant thereafter. Of current inputs to the available nutrient compartment of the forest ecosystem, 50% is from atmospheric bulk deposition, 24% from net soil release, 11% from dry deposition, 11% from root exudates and 4% is from canopy leaching. Comparing ecosystem processes for S from 1964-1969 to 1993-1998, atmospheric bulk deposition decreased by 34%, stream output decreased by 10%, net annual biomass storage decreased by 92%, and net soil release increased by 184% compared to the 1964-1969 values. These changes are correlated with decreased emissions of SO2 from the source area for the HBEF. Average, annual bulk deposition inputs exceeded streamwater outputs by 160.0 ± 75.3 SD mol S/ha-yr, but average annual net ecosystem fluxes (NEF) were much smaller, mostly negative and highly variable during the 34 yr period (-54.3 ± 72.9 SD mol S/ha-yr; NEF range, +86.8 to -229.5). While several mechanisms may explain this small discrepancy, the most likely are net desorption of S and net mineralization of organic S largely associated with the forest floor. Our best estimates indicate that additional S from dry deposition and weathering release is probably small and that desorption accounts for about 37% of the NEF imbalance and net mineralization probably accounts for the remainder (∼60%). Additional inputs from dry deposition would result from unmeasured inputs of gaseous and particulate deposition directly to the forest floor. The source of any unmeasured S input has important implications for the recovery of soils and streams in response to decreases in inputs of acidic deposition. Sulfate is a dominant contributor to acid deposition at HBEF, seriously degrading aquatic and terrestrial ecosystems. Because of the strong relation between SO2 emissions and concentrations of SO42- in both atmospheric deposition and stream water at HBEF, further reductions in SO2 emissions will be required to allow significant ecosystem recovery from the effects of acidic deposition. The destruction or removal of vegetation on experimental watershed-ecosystems at HBEF resulted in increased rates of organic matter decomposition and nitrification, a lowering of soil and streamwater pH, enhanced SO42- adsorption on mineral soil and smaller concentrations and losses of SO42- in stream water. With vegetation regrowth, this adsorbed SO42- is released from the soil, increasing concentrations and fluxes of SO42- in drainage water. Streamwater concentration of SO42- and gross annual output of SO42-/ha are essentially the same throughout the Hubbard Brook Valley in watersheds varying in size by about 4 orders of magnitude, from 3 to 3000 ha.

Original languageEnglish (US)
Pages (from-to)235-316
Number of pages82
Issue number3
StatePublished - Sep 2002


  • Acidic deposition
  • Atmospheric deposition (wet and dry)
  • Forest disturbance
  • Forest ecosystem
  • Landscape patterns
  • Net ecosystem flux
  • Net hydrologic flux
  • SO emissions
  • Soil chemistry
  • Streamwater chemistry
  • Sulfur biogeochemistry

ASJC Scopus subject areas

  • Environmental Chemistry
  • Water Science and Technology
  • Earth-Surface Processes


Dive into the research topics of 'The biogeochemistry of sulfur at Hubbard Brook'. Together they form a unique fingerprint.

Cite this