The Bear Brook Watershed in Maine: A Paired Watershed Experiment

The Bear Brook Watershed Manipulation in Maine is a paired watershed experiment. Monitoring of the paired catchments (East Bear Brook — reference; West Bear Brook — experimental) began in early 1987. Chemical manipulation of West Bear Brook catchment began in November 1989. Process studies on the watershed, outflow observation and monitoring, and modeling simulations of predicted versus observed response, are yielding a wealth of information about the behavior of the paired catchments and their responses to the deposition of acidifying substances. Results from the studies are providing important information relevant to national policies on emissions controls.

The Bear Brook Watershed Manipulation project in Maine is a paired calibrated watershed study funded by the U. S. EPA. The research program is evaluating whole ecosystem response to elevated inputs of acidifying chemicals. The project consists of a 2.5 year calibration period (1987–1989), nine years of chemical additions of (NH₄)₂SO₄ (¹⁵N- and ³⁴S-enriched for several years) to West Bear watershed (1989–1998), followed by a recovery period. The other watershed, East Bear, serves as a reference. Dosing is in six equal treatments/yr of 1800 eq SO₄ and NH₄/ha/yr, a 200% increase over 1988 loading (wet plus dry) for SO₄ and 300% for N (wet NO₃ + NH₄). The experimental and reference watersheds are forested with mixed hard- and softwoods, and have thin acidic soils, areas of 10.2 and 10.7 ha, and relief of 210 m. Thin till of variable composition is underlain by metasedimentary pelitic rocks and cale-silicate gneiss intruded by granite dikes and sills. For the period 1987–1995, precipitation averaged 1.4 m/yr, had a mean pH of 4.5, with SO₄, NO₃, and NH₄ concentrations of 26, 14, and 7 μeq/L, respectively. The nearly perrenial streams draining each watershed have discharges ranging from 0 (East Bear stops flowing for one to two months per year) to 150 L /sec. Prior to manipulation, East Bear and West Bear had a volume weighied annual mean pH of approximately 5.4, alkalinity = 0 to 4 μeq/L, total base cations = 184 μeq/L (sea-salt corrected = 118 μeq/L), and SO₄ = 100 to 111.μeq/L. Nitrate ranged from 0 to 30 μeq/L with an annual mean of 6 to 25 μeq/L; dissolved organic carbon (DOC) ranged from 1 to 7 mg/L but was typically less than 3. Episodic acidification occurred at high discharge and was caused by dilution of cations, slightly increased DOC, significantly higher NO₃, and the sea-salt effect. Depressions in pH were accompanied by increases in inorganic Al. The West Bear catchment responded to the chemical additions with increased export of base cations, Al, SO₄, NO₃, and decreased pH, ANC, and DOC. Silica remained relatively constant. Neutralization of the acidifying chemicals occurred dominantly by cation desorption and mobilization of Al.

Bear Brook Watershed in Maine (BBWM) consists of a pair of research watersheds, East Bear Brook (EBB) and West Bear Brook (WBB). Years of research and observations have shown both watersheds have high similarity in geographic and hydrologic characteristics; a simple comparison of hydrographs from these two watersheds further substantiates this similarity.

The Object Watershed Link Simulation (OWLS) model was developed and used to simulate the hydrological processes within the BBWM. The OWLS model is a 3-dimensional, vector-based, visualized, physically-based, distributed watershed hydrologic model. Simulation results not only provide a close examination of hydrologic processes within a watershed, but also dynamically visualize the processes of flow separations and Variable Source Areas (VSA).

Results from flow separations suggest that surface flow from riparian area is the predominate component for the flood rising limb and that macropore flow from riparian area dominates during the falling limb. Soil matrix flow has little effect during flood period but is a persistent contributor to base flow. Results from VSA visualization demonstrate 3-D dynamic changes in surface flow distribution and suggest that downstream riparian areas are the major contributing area for peak flow.

As water chemistry is highly relevant to the flow paths within a watershed, simulations have provided valuable information about source of stream flow and the water migration dynamics to support the study of watershed chemistry in the BBWM. More specific linkages between the chemistry behavior and the dynamic hydrologic processes should become the next simulation effort in the watershed study.

There are many questions that are critical to watershed chemistry studies like: which flow component (surface flow, macropore flow, soil matrix flow) predominates during peak flows? How do the flow components distribute during a flood event? How do flow contributions differ between these two watersheds? Which portion of the watershed contributes the most to the peak flows? These questions remain unknown from previous observations and only can be addressed with a physically-based distributed model.

The Bear Brook Watershed in Maine, USA is the site of a paired watershed study. West Bear (WB) catchment is being artificially acidified with 1,800 eq ha⁻¹ y⁻¹ of (NH₄)₂SO₄. East Bear (EB) serves as the control. After six years of artificial acidification, volume-weighted concentrations in WB, normalized to EB, increased approximately as follows, in Reg L-¹: H⁺, 15; Al (umoles), 50; Al (p.eq L-¹), 100; Ca, 50; Mg, 20; Na, 10; K, 2; SO4, 120; NI-1,, 2; NO₃, 80; HCO₃ has decreased 10 paq L’. Based on changing chemistry, several inferences can be made about soil-soil water interactions.

Dry (NH₄)₂S0, (1,800 eq⁻¹ h⁻¹ yr⁻¹) has been applied to the western of two contiguous 10 ha catchments at the Bear Brook Watershed in Maine (BBWM) since November, 1989. The initial rapid and significant response in both S and N in West Bear, compared to the reference East Bear, slowed after three years. Annual S retention of the total experimental treatment decreased from 86 to 34%, with a seven year cumulative retention of 59%. Hydrology influences the export flux of S; S is retained more in dry seasons and dry years. The annual retention of N has decreased from 96 to 81%, with a cumulative retention of 82%. The export of N from the reference watershed has declined from 178 to 23 eq⁻¹ h⁻¹ yr⁻¹ during the treatment period. The treatment N (as NH₄) initially stimulated nitrification, and caused pre-existing N to be lost in runoff, rather than the treatment N. Retention of the treatment N has decreased to approximately 80%. The majority of the retained N is stored in the soil, but the reasons for the decreased flux from the reference watershed are not known.

Phosphorus chemistry in streams was evaluated at the paired watershed study at the Bear Brook Watershed, Maine. The West Bear catchment has been treated bimonthly since 1989 with 1,800 eq (NH₄)S₂O₄ ha⁻¹ yr⁻¹. East Bear was the untreated reference watershed. During 1993, concentration of total phosphorus (P) in weekly samples from East and West Bear Brook ranged from 0 to 15 μg L⁻¹. The median values were 2 and 4 µg L⁻¹ for East and West Bear, respectively. During a high discharge event in January of 1995, the concentration of dissolved P remained relatively constant (≤ 3 μg L⁻¹) as pH decreased from 5.63 to 5.08 and from 5.14 to 4.75 in East and West Bear, respectively. The concentration of total P increased to ca. 60 μg L⁻¹ during the rising limb of the hydrograph in West Bear, four times the value in East Bear, total P then declined rapidly as discharge remained high followed by an increase. Dissolved Al increased in both streams during the episodic acidification. West Bear, the more acidic, had concentrations of dissolved Al four times those of East Bear (maximum of 1.1 mg L⁻¹ versus 0.25 mg L⁻¹). Acid-soluble particulate Al increased to 0.2 and 4.2 mg L⁻¹ for East and West Bear, respectively, in parallel to total P (but was 10² greater than total P) and then declined in parallel to total P while discharge remained high. Total P, dissolved P, and particulate Al did not relate to pH. Total P and particulate Al and Fe were strongly correlated. Concurrently, base cations remained relatively constant or decreased slightly. Particulate acid-soluble Al exceeded particulate acid-soluble base cations. We hypothesize that the particulate P was occluded in, or adsorbed on, acid-soluble particulate Al(OH)₃. This Al(OH)₃ precipitates as emerging acidic groundwater degasses CO₂ and pH rises. The export of Al and P is greater from the treated watershed because the induced acidification is translocating more Al from soils to the stream. Most of the export of P is related to acid-soluble Al particulate material

This study was conducted to determine the response of stream water DOC and organic acidity to increased inputs of ammonium sulfate to a whole catchment. Precipitation, throughfall, soil solutions (from Spodosols) and stream waters were characterized for DOC concentrations and fractions (hydrophobic acids and neutrals, hydrophilic acids, bases, and neutrals) in both the control (East Bear) and the treatment (West Bear) catchments of Bear Brook Watershed, Maine (BBWM), a northern hardwood forest. In all solutions except precipitation, DOC was composed primarily of organic acids, with hydrophobic acids dominating (> 60% of DOC) in forest floor leachates (5000 μmol C L⁻¹), and a balance of hydrophobic and hydrophilic acids in deep B horizons and stream waters (; 150 μmol C L⁻¹). Stream waters had higher concentrations of DOC during storm or snowmelt events (high discharge), often reaching 300 to 400 µcool C L⁻¹. Forest floor leachate C was rapidly attenuated by the mineral soils under all flow conditions, indicating how important mineral soil sorption of DOC was in reducing the loss of C via surface water from BBWM. No differences occurred between control and treatment streams for concentration or composition of DOC due to treatment from 1989 through 1994. In 1995, West Bear Brook had much lower concentrations of DOC than East Bear for the first time. However, this occurred during a year of record low runoff, suggesting that hydrology may have affected export of C. Average annual export of DOC from the catchments was similar (1000 to 2000 mol C ha⁻¹ yr⁻¹). Organic anions in streamwaters increased slightly during high flow events (e.g., East Bear had means of 15 and 19 μeq L⁻¹ organic anions during base flow and high discharge in 1995). Treatment of West Bear caused a decrease in organic anions, both in concentration and contribution to overall anion composition (organic anions during high discharge as a percentage of total anions decreased from about 8 to 4%, for 1987–89 and 1993–95 samples, respectively). This was probably due to decreased solution pH (greater protonation of organics) and higher concentrations of inorganic anions. Overall, there were no clear, detectable changes in stream water DOC, with only minor changes in organic anions, as a result of treatment with ammonium sulfate.

Buried mineral soil-bags and natural solutions were studied as indicators of forest ecosystem response to elevated N and S inputs at the Bear Brook Watershed in Maine (BBWM). The BBWM is the site of a paired watershed manipulation experiment in a northern New England forested ecosystem. The study includes two small (~l0 ha each) catchments dominated by northern hardwood forests with red spruce in the upper elevations. Treatments consist of (NH4)3SO4 applied to the West Bear watershed six times per year, increasing N and S deposition 3x and 2x above ambient values, respectively. Buried mineral soil-bag changes over time reflected both the native soil environment and the treatments. Most of the treatment effects on mineral soils were evident as higher inorganic S found in the treated watershed soils. Adsorbed SO, in the buried mineral soil-bags increased by approximately 40% under softwood stands and 50% under hardwood stands over the study period. Hardwood soil solutions responded with significant increases in NO₃ and SO, concentrations that resulted in accelerated cation leaching, primarily Ca and Al. Few differences that could be attributed to treatments were evident in soil solutions under softwoods. No treatment effects were evident in throughfall and stemflow chemistry.

We conducted a ¹⁵N-tracer study in a fertilized, forested catchment at the Bear Brook Watersheds in Maine (BBWM), USA, in order to characterize N cycling processes, identify sinks for ammonium-N additions, and determine the contribution of the experimental ammonium additions to nitrate exports from the treated catchment. Distributions of 15N in plant tissues, soils, precipitation and streamwater collected before adding tracers showed that nitrate-N (the dominant form of inorganic N deposition at the site) inputs under ambient conditions were depleted in 15N relative to plants and that soil was enriched in 15N relative to plants. The 15N content of streamwater nitrate was within the range of 15N contents in natural plant tissues, suggesting that nitrate deposited from the atmosphere is reduced and assimilated into soil and plant N pools before being leached as nitrate from the catchment. Variations in 15N natural abundances also suggested that most N uptake by trees is from the forest floor and that nitrification occurs in soils at this catchment under ambient conditions. Changes in 15N contents of plant tissues, soils and streamwater after adding a 15N tracer to the ammonium sulfate fertilizer applied to the treated catchment showed that soils were the dominant sink for the labeled ammonium. Surface soils (Oea horizon plus any underlying mineral soil to 5cm depth) assimilated 19 to 31 percent of the 42 kg ha⁻¹ of 15N-labelled ammonium-N during the tracer study. Aboveground biomass assimilated 8 to 17 percent of the labeled ammonium-N additions. Of the three forest types on the catchment, the soil:biomass assimilation ratio of labeled-N was highest in the spruce forest, intermediate in the beech-dominated hardwood forest and lowest in the mixed hardwood-spruce forest. Although ammonium sulfate additions led to increases in streamwater nitrate, only 2 of the 13 kg ha⁻¹ of nitrate-N exported from the catchment during the 2 years of tracer additions was derived from the 42 kg ha⁻¹ of labeled ammonium-N additions.

The difficulty of measuring ¹⁵N in dilute solutions has limited the potential of ecosystem labeling experiments and model comparisons. By concentrating the N from large (up to 20 L) water samples on ion-exchange resin columns, we obtained enough N for accurate and reproducible ¹⁵N measurement using the Teflon tape diffusion method. Analysis of standards demonstrated > 95% recovery of inorganic N from samples, and δ¹⁵N values comparable to those obtained using standard distillation methods. The value of the blank at our laboratory was 0.16 μmoles N. Analytical precision was within 2 ‰ δ¹⁵N when samples of streamwater from the Bear Brook Watersheds (1800 eq ˙ ha⁻¹ ˙ yr⁻¹ N addition at 192 ‰ δ¹⁵N) were kept frozen until analysis. The analytical process is lengthy, but once set up can easily be performed for large numbers of samples.

The paired watershed experiment at the Bear Brook Watershed in Maine (BBWM) provided an opportunity to study changes in forest soil O horizon properties as a result of experimental, chronic N additions. The West Bear Brook watershed received elevated N and S inputs since November 1989 as bimonthly applications of (NH₄)₂SO₄. Forest floor samples (O horizon) were collected in July of 1992 from three dominant stand and five soil types at BBWM. The (NH₄)₂SO₄ amendments in the treated watershed (West Bear) stimulated potential net nitrification, but significant increases were found only in hardwood O horizons after three years of treatment. Hardwood stand forest floor soil materials had the lowest C:N ratios (mean=23), compared with mixedwood (mean=27) and softwood stands (mean=33). NH₄-N accounted for over 95% of the inorganic N in the forest floor. The lack of a strong relationship between soil type and potential net N mineralization at BBWM, coupled with conflicting results in the literature, suggested that stand characteristics were more important than conventional soil nomenclature based on pedogenetic features, or 2.5 years of treatments, in defining differences in soil N dynamics and responses to increased N inputs.

Foliar chemistry was examined in mature sugar maple (Acer saccharum Marsh), red maple (Acer rubrum L.), American beech (Fagus grandifolia Ehrh.), and red spruce (Picea rubens Sarg.) in response to chronic, watershed-level additions of ammonium sulfate [(NH₄)₂SO₄]. Following four years of treatment, N concentrations were significantly higher in foliage from the treated watershed for all four species, with increases ranging from 6% in American beech to 33% in sugar maple. Sugar maple foliage from the treated watershed had significantly lower Ca concentrations (18%). Concentrations of K were significantly lower in beech (13%) and red spruce (9%) from the treated watershed. Foliar Mg was not different between watersheds. Aluminum concentrations were significantly higher in the foliage from the treated watershed for beech (18%), red maple (33%), and sugar maple (65%), but no differences in Al concentration occurred in current year red spruce foliage. Red spruce foliage resampled following a fifth year of treatment contained higher concentrations of N and Al and lower concentrations of Ca and Mg in the treated watershed. Despite these differences in red spruce foliar chemistry, wood production and density did not appear to be affected by the treatment.

Differences in the foliar chemistry between the treated and untreated watershed may reflect the temporal and spatial integration of changes taking place in the soil of the treated watershed. Increased N is likely directly due to the N contained in the (NH₄)₂SO₄ treatment. Labile Ca and other cations in the treated watershed would be expected to initially increase and then decrease in response to the treatment, with these changes beginning at the top of the forest floor. Thus, lower cation concentrations in foliage from the treated watershed may reflect the fact that cations in the uppermost portions of the soil were rapidly depleted, even though deeper soil layers were experiencing increased Ca release due to cation exchange effect of the acidification. The generally higher Al in foliage from the treated watershed is likely due to the mobilization of inorganic Al in the soil as has been reported previously for the treated watershed. Collectively these results suggest that the long-term deposition of acidifying substances containing N and S not only influence the cycling of N within these systems, but may also alter the cycling of other important nutrients and Al.