Soil Solution Phosphorus Dynamics in a Whole-Tree-Harvested Northern Hardwood Forest

Forest harvest drastically alters nutrient cycling, yet stream water and mineral-soil solutions typically show little change in P concentrations after forest disturbance. Changes in P availability and movement are more likely to be detected in forest-floor solutions, since P can be strongly sorbed in soil. Tension-free lysimeters were used to compare forest-floor and mineral-soil solution total-P concentrations in whole-tree-harvested and undisturbed forest sites for 2 yr after harvest in the Hubbard Brook Experimental Forest in New Hampshire. Mean total-P concentrations in solution in the forest floor (Oa horizon) were significantly higher in whole-tree-harvested sites (248 Mg P L") than in undisturbed forest sites (67 Mg P L-'), presumably due to reduced plant uptake of P. Mean P concentrations in mineral-soil solution remained low after whole-tree harvest (32 Mg P Lin the Bh and 15 Mg P L" in the Bs), consistent with the high P-sorption capacity of the mineral soil. Phosphorus flux in soil solution from the Oa to the B horizon, calculated from monthly mean P concentrations and monthly water flux simulated by a hydrologic model, was greater in whole-tree-harvested sites (1.0 kg P hayr-) than in undisturbed forest sites (0.3 kg P hayr-) in the first 2 yr after the disturbance. This redistribution of P from the forest floor to the mineral soil far exceeds stream-water export of dissolved P (0.02 kg P ha' yr) but is small compared with forest-floor stores (85 kg P ha-). T WHOLE-TREE HARVEST of Watershed 5 (W-5) at the Hubbard Brook Experimental Forest (HBEF) in New Hampshire provided an opportunity to study the effects of forest disturbance on P cycling within the ecosystem (Yanai, 1990). Phosphorus concentrations in stream water change little after forest harvest at this or at other sites (Hobbie and Likens, 1973; Likens and Bermann, 1974; Douglas and Swank, 1975; Fredriksen et al., 1975), in contrast to N concentrations, which can rise dramatically (Pierce et al., 1972; Corbet et al., 1978; Bormann and Likens, 1979; Vitousek and Melillo, 1979). Soil solutions have also failed to show increased P concentrations after forest cutting (Richardson and Lund, 1975; McCpll, 1978), despite the drastic alteration of nutrient cycling caused by deforestation. Both the low concentrations of P observed in stream water and soil solutions and the lack of response to clear-cutting can be explained by P sorption in the soil (Wood et al., 1984). Since the forest Boyce Thompson Inst, for Plant Research, Cornell Univ., Ithaca, NY 14850. Research conducted at Yale School of Forestry and Environmental Studies, New Haven, CT 06520. Received 7 Nov. 1990. *Corresponding author. Published in Soil Sci. Soc. Am. J. 55:1746-1752 (1991). floor lacks this P-sorption capacity, solution concentrations in the forest floor are more likely to display changes in P availability and movement following forest disturbance than are mineral-soil solutions or stream water. In the undisturbed HBEF, biological cycling of P is most active in the well-developed forest floor (Wood et al., 1984). Decompose]organisms, absorbing roots, and organic-P input via litterfall and fine-root turnover are concentrated in the Oa horizon. Because P uptake by vegetation is severely reduced immediately following whole-tree harvest, while decomposition and P mineralization continue and may even accelerate (Bormann and Likens, 1979), concentrations of P in soil solution in the forest floor were expected to increase after harvest. Accordingly, more P was expected to leach from the forest floor into mineral soil in harvested sites than in undisturbed forest. Phosphorus concentrations in mineral-soil solutions were expected to remain low, due to the high P-sorption capacity of the B horizons. Increased P loss from the forest floor as a result of forest disturbance could represent an important perturbation of the P cycle, since the undisturbed forest relies heavily on the forest floor for P supply (Yanai, 1990). Although P adsorption retains most of the translocated P in the mineral soil, mineral-soil P is less available to plants than is forest-floor P (Hoyle and Bjorkbom, 1969; Wood, 1980). Therefore, accelerated loss of P from the forest floor to the mineral soil could have important consequences for the supply of P for future forest gro\vth. To test the predictions regarding soil solution-P concentrations, tension-free lysimeters were used to compare total P concentrations in solution in the forest floor and mineral soil in whole-tree-harvested and undisturbed small-watershed ecosystems for 2 yr after harvesting. To quantify P loss from the forest floor following disturbance, leaching of total dissolved P through the monitored soil horizons was calculated from P concentrations and from water flux simulated by a hydrologic model. These calculated P transfers were used in constructing P budgets of the undisturbed forest and of the recovering forest in the first 2 yr postharvest (Yanai, 1990).