No abstract

Abstract. A synthesis of the biogeochemistry of K was conducted during in the reference and human-manipulated watershed-ecosystems of the Hubbard Brook Experimental Forest (HBEF), NH. Results showed that during the first two years of the study , which coincided with a drought period, the reference watershed was a net sink for atmospheric inputs of K. During the remaining years, this watershed has been a net source of K for downstream ecosystems. There have been long-term declines in volume-weighted concentration and flux of K at the HBEF; however, this pattern appears to be controlled by the relatively large inputs during the initial drought years. Net ecosystem loss (atmospheric deposition minus stream outflow) showed an increasing trend of net loss, peaking during the mid-1970s and declining thereafter. This pattern of net K loss coincides with trends in the drainage efflux of SO: and NO;, indicating that concentrations of strong acid anions may be important controls of dissolved K loss from the site. There were no long-term trends in streamwater concentration or flux of K. A distinct pattern in pools and fluxes of K was evident based on biotic controls in the upper ecosystem strata (canopy, boles, forest floor) and abiotic controls in lower strata of the ecosystem (mineral soil, glacial till). This biological control was manifested through higher concentrations and fluxes of K in vegetation, aboveground litter, throughfall and forest floor pools and soil water in the northern hardwood vegetation within the lower reaches of the watershedecosystem, when compared with patterns in the high-elevation spruce-fir zone. Abiotic control mechanisms were evident through longitudinal variations in soil cation exchange capacity (related to soil organic matter) and soil/till depth, and temporal and disturbance-related variations in inputs of strong-acid anions. Marked differences in the K cycle were evident at the HBEF for the periods 1964-69 and 1987-92. These changes included decreases in biomass storage, net mineralization and throughfall fluxes and increased resorption in the latter period. These patterns seem to reflect an ecosystem response to decreasing rates of biomass accretion during the study. Clearcutting disturbance resulted in large losses of K in stream water and from the removal of harvest products. Stream losses occur from release from slash, decomposition of soil organic matter and displacement from cation exchange sites. Elevated concentrations of K persist in stream water for many years after clearcutting. Of the major elements, K shows the slowest recovery from clearcutting disturbance.

We investigated changes in soil‐atmosphere flux of CH4, N2O, and NO resulting from the succession of pasture to forest in the Atlantic lowlands of Costa Rica. We studied a dozen sites intensively for over one year in order to measure rates and to understand controlling mechanisms for gas exchange. CH4 flux was controlled primarily by soil moisture content. Soil consumption of atmospheric CH4 was greatest when soils were relatively dry. Forest soils consumed CH4 while pasture soils which had poor drainage generally produced CH4. The seasonal pattern of N2O emissions from forest soils was related exponentially to soil water‐filled pore space. Annual average N2O emissions correlated with soil exchangeable NO3− concentrations. Soil‐atmosphere NO flux was greatest when soils were relatively dry. We found the largest NO emissions from abandoned pasture sites. Combining these data with those from another study in the Atlantic lowlands of Costa Rica that focused on deforestation, we present a 50‐year chronosequence of trace gas emissions that extends from natural conditions, through disturbance and natural recovery. The soil‐atmosphere fluxes of CH4 and N2O and of NO may be restored to predisturbance rates during secondary succession. The changes in trace gas emissions following deforestation, through pasture use and secondary succession, may be explained conceptually through reference to two major controlling factors, nitrogen availability and soil‐atmosphere diffusive exchange of gases as it is influenced by soil moisture content and soil compaction.

A systematic examination of nitrogen cycling in disturbed forest ecosystems demonstrates that eight processes, operating at three stages in the nitrogen cycle, could delay or prevent solution losses of nitrate from disturbed forests. An experimental and comparative study of nitrate losses from trenched plots in 19 forest sites throughout the United States suggests that four of these processes (nitrogen uptake by regrowing vegetation, nitrogen immobilization, lags in nitrification, and a lack of water for nitrate transport) are the most important in practice. The net effect of all of these processes except uptake by regrowing vegetation is insufficient to prevent or delay losses from relatively fertile sites, and hence such sites have the potential for very high nitrate losses following disturbance.

The controls of potential nitrogen mineralization, nitrate production, and nitrate mobilization in a wide range of forest ecosystems were investigated through a combination of field and laboratory experiments. Trenched plot experiments were performed in 17 forests, and laboratory incubation studies of potential ammonium and nitrate production were made on soils from 14 of these sites. The site with the greatest potential for nitrate production in the laboratory was a New Hampshire northern hardwoods forest. Several other sites, including New Hampshire balsam fir, Indiana maple—beech, New Mexico aspen, and Oregon western hemlock forests, also had high potential nitrate production. All of these sites also had rapid nitrate movement to below the rooting zone following trenching in the field. Of nine processes which could be important in preventing or delaying solution losses of nitrate from disturbed forests, two appeared most important among the forests we examined. Low net nitrogen mineralization (caused by either nitrogen immobilization or low gross nitrogen mineralization) and lags in nitrification (probably caused by either low initial populations of nitrifying bacteria or the allelochemic inhibition of nitrification) were identified as important in several sites and in different regions. A direct relationship between the amount of nitrogen in annual litterfall and the proportion of forest floor nitrogen mineralized in laboratory incubations was observed, suggesting that refractory organic nitrogen compounds are produced in nitrogen—poor sites. An inverse relationship was found between the amount of nitrogen in litterfall in these and other sites and the carbon:nitrogen ratio of that litterfall, suggesting that the immobilization capacity of litter is increased in nitrogen—poor sites. The presence and length of lags in nitrification were inversely correlated with the mean concentration of mineral nitrogen in mineral soil. These patterns suggest that nitrogen retention within disturbed forest ecosystems can be caused by low nitrogen availability prior to disturbance.