DIRT Litter Manipulation Experiment at Harvard Forest since 1990

The DIRT Experiment (Detritus Input and Removal Treatments) is a long-term study of controls on soil organic matter formation. Our goal is to assess how rates and sources of plant litter inputs control the accumulation and dynamics of organic matter and nutrients in forest soils over decadal time scales. Results from 11 years of field and laboratory studies demonstrate the relative importance of above- and belowground sources on soil organic matter (SOM) dynamics and show emerging long-term non-linear changes in soil carbon release and storage. Treatments established in a mixed hardwood stand in 1990 are: doubling annual aboveground litter (DL), exclusion of aboveground litter (NL), exclusion of root inputs by trenching (NR), and exclusion of aboveground litter and root inputs (NI), on replicated 3m x 3m plots (n=3 for treatments, 6 for controls). The O/A-less treatment, implemented in 1991, tracks the recovery of impoverished soil by replacing O and A horizon soil with B horizon material and allowing normal litter inputs thereafter. Comparison of data among treatments (soil respiration, soil solution chemistry, soil physical and chemical properties, and microfaunal and microbial community structure) allows us to determine the contributions of live roots, above-ground litter, and belowground detritus to SOM and nutrient dynamics in this forest soil. Similar experiments in Pennsylvania, Wisconsin, Oregon, and Hungarian forests provide information on these processes across climate and soil texture gradients. First-year soil respiration results from the Harvard Forest DIRT plots showed that live root respiration, production of aboveground litter (leaf, twig, other fine litter) and fine root detritus each constitute about one-third of C inputs to soil. Soil respiration is influenced more by root inputs than aboveground litter in this forest. CO2 efflux from root-excluded soils (NR, NI) declined to 32% of controls over the first 11 years of treatments as soil C became more recalcitrant. Excluding aboveground inputs (NI) had little additional negative effect when root inputs were excluded (NR). Doubling or excluding aboveground litter proportionally increased or decreased respiration, respectively, from root-intact soils during the first four years. However, after year 8, DL respiration dropped to near or below control levels, indicating decreasing decomposition rates. O/A-less soil respiration rates in year 11 surpassed the root-excluded treatments. C and N stocks respectively increased or decreased in response to doubling or excluding aboveground litter inputs to root-intact soil. However, stocks declined only slightly in NI soils relative to NR. Changes in soil respiration relative to total soil C mass indicate changing soil organic matter (SOM) quality across treatments and over time. Root inputs remain the stronger influence on the proportion of respired labile C, which declined to 40% of control levels in the root-free soils. Results suggest that aboveground inputs more strongly influence SOM mass, but root inputs have a stronger effect on SOM quality. 10 years of doubling litter inputs increased total soil C but by year 10 respired labile C declined below controls, demonstrating a long-term negative effect on decomposition. Total C in O/A-less soil did not increase markedly but the large labile proportion dropped by half as the young organic matter became more stabilized after 10 years of decomposition. Soil respiration response to soil temperature was strongly and significantly influenced by treatment, and treatment effects increased with time. Respiration by fine roots and associated rhizosphere organisms was more responsive to temperature than was bulk soil respiration. DL Q10 fell steadily from year 4 to year 11 as decomposition rates declined, and was below NL Q10 by year 10. NL Q10 changed little over time. NI Q10 decreased more than NR despite the fact that its respiration rates are similar to or greater than NR. Q10 for NI and NR decreased more over time than any other treatment. O/A-less soil Q10 increased over time, and was higher than the root-excluded treatments’ Q10 in year 11. Changes in soil solution organic chemistry required at least 5 years of litter manipulation. By year 7, forest floor DOC concentrations were significantly higher in DL plots and significantly lower in O/A-less plots. Response was mixed in the litter and/or root-excluded treatments. Mineral soil solution chemistry was not affected by treatments. The Oea horizon contributed 40% to forest floor soil solution DOC, leaf litter 44%, and root exudate and decay 16%. Roots appeared to be a sink for DON. Oea soil appears to contribute 107% and leaf litter 39% to forest floor soil solution DON. Mean annual DOC flux was strongly related to forest floor C:N ratio but DON flux was not. Mean annual DOC and DON fluxes were positively related to fungal biomass, suggesting that fungal biomass may be responsible for a large proportion of DOC and DON production. Seasonal changes in DOC:DON ratios in root-intact treatments DL, NL, and controls suggest a decoupling of DOC and DON production. Treatment patterns of DOC flux in year 7 correspond to cumulative DOC and CO2 release from incubated organic horizon soils collected in year 5, as did CO2 efflux measured in the field in year 8. DOC losses from the incubated forest floors were 10 percent of CO2-C gas losses. Total DIN mineralized in the incubated soils collected in year 5 was also influenced by treatments. Lack of net N mineralization response to variations in aboveground litter suggests that microbial immobilization exerts a strong control over soil N dynamics. Although net N mineralization rates were lower in soils from root-excluded plots, nitrification rates were much higher. Incubated soils from the Pennsylvania site showed similar results. These findings suggest that the absence of roots and mycorrhizal hyphae favor nitrifying bacteria. After 5 years of treatments, microflora appeared to follow patterns of carbon availability and recalcitrance. Data suggest that fungal:bacterial ratios decline with increasing recalcitrance of soil carbon. Total fungal biomass tended to follow patterns of C and N content, being highest in DL and lowest in NL and NI. Fungal biomass was much greater than bacterial biomass in all treatments, yet aboveground litter inputs may be more important substrates for fungi than are roots. Treatments did not affect the active biomass of fungi or bacteria in forest floors. That litter manipulations strongly affected N mineralization and respiration in laboratory incubated soils suggests that the activities of microbial functional types were influenced by treatments. However, microbial populations are poor predictors of process rates. In summary, results from field data and intersite soil incubations suggest that aboveground inputs exert a stronger influence on SOM mass, but root inputs have a stronger effect on its quality. The pool of turning over N is slow, having a different dynamic than the faster mineralizable C pool. Metabolism of roots and rhizosphere organisms is more temperature-sensitive than bulk soil organisms. Exclusion of roots had a greater effect on microbial processes than either doubling or excluding aboveground inputs. Declining decomposition rates accompanied by increasing soil C in the DL treatment suggests long-term non-linear changes in soil microbial activity which could lead to increased long-term soil C storage beyond expectations. Changes in above- and belowground plant inputs and their influence on temperature-controlled processes will be significant in determining the effects of a warmer world on the net flux of carbon from soils to the atmosphere.