Date of Award
Doctor of Philosophy
Ecology and Evolutionary Biology
Frank W. Woods
Gary S. Henderson, Edward E. C. Clebsch, D. L. Bunting
Aspects of the organic carbon cycle in a deciduous forest watershed were studied with emphasis on meteorological inputs to the watershed system, transfers of organic material from the terrestrial to the aquatic subsystems, the hydrologic transport of organic carbon in the stream subsystem, and organic outputs from the watershed system via geological vectors.
Transfer of large (> 1 mm) particulate organic material to the heterotrophically based aquatic subsystem occurred primarily as direct leaf fall and blow-in. Annual litterfall was 222.1 g/m2 organic carbon, of which leaves contributed 80.6%, fruits and reproductive parts 11.3%, twigs 5.6% and frass 2.5%. Peak inputs for all components except frass occurred during the autumn.
The steepest southwest-facing slopes generally contributed the highest inputs, with peak inputs for leaf and twig material occurring during the winter to early spring period when aeolian forcing factors were greatest. Fruit and frass inputs were more closely related to seasonal litterfall patterns than to aeolian factors. Total yearly organic carbon input of blow-in to the stream channel was 52.1 g/m2, of which 82.4% was leaves, 13.6% was fruits, 3.8% was twigs, and 0.2% was frass.
Standing crops of large ( > 1 mm) particulate organic material in four stream habitat types (dry gravel > stream gravel > dry bedrock > bedrock pools) showed significant date and habitat type differences, reflecting the influences of the hydrologic cycle and within-system biological processing on the allochthonous inputs. Largest standing crops were found during November (62.76 - 283.00 g/m2) and smallest in early February (0.55 - 13.21 g/m2), the decrease due mainly to hydrologic transport. Largest relative contribution of leaves occurred during autumn (> 90% in November), while lowest relative contribution occurred during the June sampling, due mainly to increased twig inputs during the previous weeks.
Concentrations of dissolved organic carbon (DOC) and fine particulate organic carbon (FPOC) in throughfall were highest in July (32.32 and 20.27 mg/1, respectively) and lowest in December (1.90 and 0.49 mg/1, respectively), when canopy was absent and rainfall heavy. Both DOC and FPOC data had a negative logarithmic relationship with volume of throughfall. Similar trends were seen for DOC and FPOC in incident precipitation, but concentrations were generally lower (1.55 - 8.78 mg/1 for DOC and 0.88 - 2.12 mg/1 for FPOC), as were r2 values. Trends for concentrations of DOC and FPOC attributable to canopy removal generally followed those for throughfall, ranging from 0.67 to 23.46 mg/1, and 0.19 to 18.18 mg/1, respectively). For throughfall, the presence or absence of a canopy generally influenced both the slope and intercept of the regressions. Inputs of organic carbon in throughfall and incident precipitation showed less month to month variation than concentration, with only FPOC in incident precipitation showing a significant relationship (possibly logarithmic) to volume of rainfall. DOC inputs ranged from 3.32 to 10.05, 1.40 to 6.55, and 0.08 to 8.64 kg/ha/mo for throughfall, incident precipitation, and canopy contribution, respectively. For FPOC, inputs ranged from 0.85 to 3.82, 0.55 to 3.18, and -0.19 to 3.90 kg/ha/mo for throughfall, incident precipitation and canopy contribution, respectively. Two-thirds of the total input (95.2 kg/ha/yr) of organic carbon throughfall was in dissolved form. Net canopy removal for the year was 38.68 kg/ha, of which 61.6% was dissolved.
Soil water had low concentrations of DOC for all four forest types studied (0.62 - 2.60 mg/1). Significant date and forest type differences were apparent, with highest concentrations in late summer and fall through winter and lower concentrations in spring and early summer. Oak-hickory and chestnut oak types had higher concentrations of DOC than the pine and yellow poplar types due, in the case of pine, to a lower pool of decomposable and/or soluble organic material and, in the latter case, to lower slope position, allowing longer contact of infiltrate with the soil matrix. No significant relationship was found between concentration of DOC volume of infiltrating rainfall. Outputs from the computer code PROSPER were utilized to calculate the flux of organic material (1.34 g/m2 /yr) past the 75 cm soil depth.
Springflow concentrations of DOC and FPOC were low and invariable through the year, with highest concentrations (0.59 and 0.82 mg/1, respectively, for DOC and FPOC) recorded during the week which included the most intense storm of the study period. Most weekly means were below 0.25 mg/1 for both species, due primarily to the diffuse flow nature of the spring system, allowing intimate contact of the water with the bedrock system.
Analyses of storm cycles occurring throughout the water year indicated that DOC followed the hydrograph much more closely than FPOC, the latter responding very quickly to hydrographic rises, with concentrations often decreasing before peak discharge was reached. Ratio of concentration of DOC to concentration of FPOC thus decreased early in the storm, increasing after the initial FPOC peak. For "non-rain" or delayed increases in streamflow, FPOC concentration behaved similarly to that for a rise due to direct channel input, but DOC concentration showed no increase. The different responses were due to different source areas of streamflow and organic material. Highest concentrations of FPOC (77 mg/1) were seen during major storms, while for DOC, highest concentrations (13.55 mg/1) were found in the summer to fall period due to throughfall contributions. Concentrations of DOC and especially FPOC were dependent on antecedent events on the watershed. Regression analyses for the relation of DOC and FPOC concentrations in individual samples to discharge showed poor fits, especially for DOC, with the fits improving somewhat when data were sorted by season and month. For the same regression form, FPOC had the greater slope, and the ratio of DOC/FPOC vs discharge virtually always had a negative slope, further indicating the dominance of FPOC at high discharge. Good fits were found for outputs of DOC and FPOC vs discharge, with the untransformed regression best explaining the DOC output trends and the logarithmic relationship best fitting the FPOC data. Slopes for FPOC were generally higher, with highest slopes for both DOC and FPOC in summer and fall. Generally similar results were found for regressions involving the weighted weekly concentration and weekly output and discharge, except that the semilogarithmic relationship gave the best fit in many cases.
Baseflow concentrations were very low (0.1 -0.4 mg/1) for both DOC and FPOC, with only the autumn litterfall period showing elevated DOC baseflow levels. Lowest weekly outputs occurred during the late summer - early all period due to low baseflow and lack of significant storm events. Low baseflow levels led to large ranges (greater than two orders of magnitude) in concentration during the year.
Highest weekly weighted concentrations (1.87 mg/1 for DOC and 7.45 mg/1 for FPOC) and lowest ratios of DOC to FPOC occurred during week 48 of 1973, during which the highest peak flow of the study occurred. Large storms during other parts of the year, while generally decreasing in the DOC/FPOC ratios, did not decrease the ratio to the extent that the storm of week 48 did, due to less standing crop of FPOC in and adjacent to the stream channel. The forty-eighth week, with 9% of the discharge for the water year, had 49.6% of the total organic output for the year. For DOC this was 30.7% of the yearly output and for FPOC 58.18%. The weighted mean concentration of the 1973-1974 water year was 0.56 mg/1 DOC and 1.16 mg/1 FPOC. Total output for the 1973-1974 water year was 1016.9 kg, of which 32.5% was DOC and 67.5% was FPOC.
Litterfall dominated the inputs of organic carbon to both the forest floor and the stream system. DOC losses from the 75 cm layer were inconsequential compared to losses of CO2. Loss of TOC i streamflow, corrected for interbasin transfer, was 24.89 kg/ha/yr, amounting to only 44% of the meteorological input of DOC and FPOC. However, the two organic types behaved quite differently with FPOC actually showing a net loss of material for the year.
Comiskey, Charles Edward, "Aspects of the Organic Carbon Cycle on Walker Branch Watershed: A Study in Land/Water Interaction. " PhD diss., University of Tennessee, 1978.