Effects of a Massive Ice Storm on the Northern Hardwood Forest and Associated Aquatic Ecosystems in the Hubbard Brook Valley and Surrounding White Mountain National Forest Region

        A massive ice storm struck northern New England and southeastern Canada during 7-9 January 1998. This storm was devastating to power and telephone transmission lines and had great impact on vegetation, particularly the deciduous forest at higher elevations (>500 m). The storm has been described as the worst of this type in a century (e.g. The New York Times, 13 January 1998). Many have described it as the worst forest disturbance since the hurricane of 1938, which caused extensive forest damage throughout New England (see Bormann and Likens 1979).

        The vegetation of the Hubbard Brook Experimental Forest (HBEF) was damaged by this storm, primarily at higher elevations where our long-term, experimental watersheds are located, including our long-term biogeochemical reference watershed -- W6. Photos and a narrative description of the disturbance can be found at www.hubbardbrook.org/yale/misc/beech98.htm.

        Damage was patchy with areas of 100% canopy loss, interspersed with areas of moderate (e.g. 25% canopy loss), to areas with little damage. The severity of damage appears to vary among species in the order, beech (Betula alleghaniensis) > yellow birch (Fagus grandifolia) > sugar maple (Acer saccharum).

        A fundamental challenge in ecology is to separate the effects of natural disturbance from those caused by anthropogenic disturbance on the structure, function, and development of ecosystems. The damage caused by this ice storm provides a unique opportunity to address this important ecological issue. A small grant for exploratory research (SGER) from the National Science Foundation is helping to fund some of the studies described below

        Based on our long-term studies of the response of the northern hardwood forest ecosystem to disturbance by various forestry practices and biological agents, we hypothesized that the disturbance from the ice storm will: (1) increase soil temperature due to increased solar radiation at the soil surface because of loss of canopy; (2) increase inputs of relatively labile carbon to the forest floor, in the form of dead roots and aboveground detritus, thereby stimulating mineralization of carbon and mineral nutrients; (3) decrease water and nutrient uptake by roots due to root death and reduction of aboveground water and nutrient demand; (4) increase yield of water as a result of reduced transpiration; and (5) increase nitrification and consequently increase leaching of solutes especially NO₃^- and K⁺. Unlike a commercially clear-cut forest, the intensity of the disturbance associated with the ice storm was highly variable across the landscape, and we anticipated that these responses will vary with the scale and intensity of canopy disturbance. A focus of our Hubbard Brook LTER (HBR-LTER) project is to assess spatial patterns of biogeochemical processes across the forest landscape. Because of the spatial heterogeneity of the ice storm disturbance, this event provided a unique opportunity to test our basic understanding of biogeochemical controls on surface water chemistry at HBEF. A prompt, detailed, quantitative assessment of the intensity, scale and arrangement of canopy disturbance will allow us to seize this opportunity.

        We would expect, based on our previous studies of forest disturbance (e.g. Likens et al. 1970; Bormann and Likens 1979; Martin et al. 1986), that NO₃^- output in stream water from these damaged forests will increase beginning in mid summer 1998. Long-term biogeochemical monitoring in W6 (HBR-LTER) will enable us to detect the overall and long-term changes, but this natural disturbance event provides an opportunity to determine how major forest damage of this type affects nutrient cycling across local, watershed and regional scales related to this disturbance event. We hypothesize further that the effects of the massive ice storm will result in marked increases in the leaching of NO₃^- and K⁺ in drainage water locally and regionally. Determining how this large-scale disturbance interacts with the effects of atmospheric deposition of nitrogen to influence regional patterns of surface water nitrogen flux is also a principal objective of this research.

        We are: (1) documenting and evaluating quantitatively the structural damage to the forest throughout the Hubbard Brook Valley and the upper catchments of the larger Pemigewasset-Merrimack basin, (2) repairing existing measurement plots and establishing new plots in disturbed and reference areas along an elevational transect west of W6, and (3) sampling and analyzing surface water chemistry across the region. We will concentrate "on-the-ground" measurements in W6 and on our adjacent (west of W6) intensive measurement plots, as well as in plots in W1 and W4. We will quantify in detail the vegetation damage on W1, 4, 5 and 6, and the local biogeochemical responses to comparable damage in the intensive plots west of W6 and in W1. This information will be combined to predict the responses of W6, in terms of longitudinal and temporal chemistry patterns as well as overall chemical mass balance. The test of these predictions should provide new insights into controls of hydrological and biogeochemical cycles in the northern hardwood forest landscape. In addition, we will predict and test surface-water chemical responses across the upper basin of the Pemigewasset-Merrimack River, using damage assessments being developed by the USDA Forest Service’s Forest Inventory and Analysis Team (FIA).

        Because of the infrastructure of the HBR-LTER, we are in an excellent position to quantify the response of soil temperature, vegetation, microbial biomass and activity, trace gas production and drainage water amount and chemistry from the effects of the ice storm disturbance. Our current monitoring program includes: a complete inventory of vegetation of W6, a network of plots for measurement of soil temperature, root activity, microbial biomass and activity, CO₂ and N₂O production and soil solution chemistry, and stream monitoring stations along an elevational transect. These include stream monitoring stations within the experimental watersheds, along the main Hubbard Brook, and within the greater Pemigewasset-Merrimack River basin.

        We finished a complete forest inventory of W6 during the summer of 1997 [www.hubbardbrook.org/yale/vegetation], and we now have resurveyed the forest during the summer of 1998 to quantify the disturbance associated specifically with the ice storm. (As part of the routine HBR-LTER effort, this inventory will be repeated in 2002.) Relatively few trees were uprooted and most of the damage was in crown breakage. Thus, an inventory of basal area alone will not be a good measure of the damage. Our plan is to assess damage, both from low-altitude aerial photographs and by on-the-ground inventories of broken limbs and trees. Exceptionally detailed tree allometry is available for this site (Whittaker et al. 1974), which will facilitate the quantification of canopy damage. Our estimates of canopy loss will be verified by direct measurements of leaf litterfall and indirect measurement of leaf area index (LAI) using the LiCor LAI-2000. Damage to and response of the sapling and shrub layer also will be assessed on sub-plots. These measurements will allow us to quantify changes in solar radiation reaching the soil surface and consequent responses in soil temperatures.

        We anticipate that belowground responses will vary markedly with the degree of LAI reduction and the scale of this reduction. Monitoring plots have been established in our intensive monitoring area west of W6 and in W1, spanning a representative range of canopy damage. Plots have been equipped for measurement of soil temperature, soil solution chemistry, litterfall, microbial activity, root activity and gas emissions, using our standard protocols. We are particularly interested in fine root responses because root death and changes in root uptake are likely to be crucial variables influencing biogeochemical responses. Root death is being examined on the basis of temporal changes in K⁺ concentration of fine roots sorted from soil (Fahey and Arthur 1994). Relative changes in root growth is being assessed in situ with horizontal root screens, a technique that provides the very large sample sizes needed to obviate spatial variability (Fahey and Hughes 1994).

        This disturbance event provides an unusual opportunity to evaluate connections between vegetation and ecosystem response, and perhaps place this intensive, but diffuse, disturbance into our concept of forest dynamics that has emphasized gaps -- a different twist on the shifting mosaic concept (Bormann and Likens 1979). Over the longer term through the ongoing HBR-LTER project, we will measure dead wood mass and ultimately its decay, soil temperature, root activity, microbial biomass and activity and soil water chemistry in plots established in this initiative.

        To assess the regional response to the ice-storm damage, we are conducting additional collection and analysis of drainage water throughout the HBEF and in downstream aquatic ecosystems. We have between 15 and 35 years of monthly and weekly baseline data to draw upon from this network. During the summer and autumn of 1998 we have sampled intensively our existing network of stream sampling stations for water chemistry. This approach includes collections within the experimental watersheds (tens of ha), as well as the Hubbard Brook Valley (~3000 ha) and throughout the Pemigewasset-Merrimack River basin (280,000 ha). As part of the HBR-LTER, we have a baseline of water chemistry from 50 sites established several years ago within the greater Pemigewasset-Merrimack basin. We are particularly interested in a NO₃^- response (see e.g. Likens et al. 1970; Likens and Bormann 1974; Bormann and Likens 1979; Martin et al. 1986).

Investigators: Steven P. Hamburg and Anne Rhoads

        Ice storms are common to the mesic forests of the northeastern United States (Boerner et al. 1988, Siccama et al. 1976). Several studies exist which have explored tree susceptibility to ice (Boerner et al. 1988, Lemon 1961, Melancon and Lechowicz 1987, Siccama et al. 1976). Others have focused on short-term changes to the affected forests (Bruederle and Stearns 1985), while others have examined long-term effects such as those on forest succession (Whitney and Johnson 1984). However, very little experimental data exist which document the recovery of the canopy following ice storms.

        In this current study, an examination of the LAI will be used to follow the recovery of the canopy of the northern hardwood forest following the January 1998 ice storm. The plots studied will be located in the middle and upper elevational range of W1 and W6. Using these watersheds will enable a comparison to be made of the canopy recovery in an untreated area (W6) with that in one where future calcium additions will be made (W1).

        Average damage codes of the dominant trees (>20 cm dbh) on the two watersheds were determined by Tom Siccama and others. Damage was characterized as a percentage of the canopy damaged (0% damage = 0,1-25% = 1, 25-50% = 2, 50-75% = 3, 75-100% = 4, crown gone = 5, uprooted or broken off at base = 6). In this study, these scores were then used to categorize each plot as to its mean damage class (0-<1, 1-<2, 2-3). The plots were further subdivided into elevational bands. After stratifying the plots by damage class and elevation, 22 plots were selected randomly from among the subcategories. Control plots are clearly needed, but there are few areas within the middle and upper elevational ranges of the watersheds that have not been affected by the storm. Therefore, 10 undamaged reference plots were selected at lower elevations on W1 and W6.

        Measurements are made using two Licor LAI 2000 instruments in remote mode. After calibrating both under identical conditions, one is used under the canopy (B reading), and the other outside the canopy (A reading). The latter acts as the necessary above-canopy measurement. The A reading is taken in an open field 2 km away from the study area. Readings will be taken at the middle and end of the growing season. All measurements are taken at either dawn or dusk.

The distribution of damage classes among plots:

Investigator: Anne Rhoads

Slide Presentation from the Annual Hubbard Brook Cooperators Meeting July 8, 1999

Slide 1 of 13

Investigator: Thomas G. Siccama

        The objective of this effort is to estimate the amounts of biomass "downed" by the ice storm. By "downed" we mean killed and eventually falling to the ground. At present a considerable amount of dead biomass is hung up in the trees as broken and hanging branches and tops. [see www.hubbardbrook.org/yale/misc/beech98.htm]

        Using these categories of damage together with allometric equations for estimating biomass of branches, we estimated the biomass killed by the ice storm. On W6 the ice storm affected about 73% of the watershed. Damage extends upward from about 2080 ft (636 m), and the following estimates are based on a per hectare estimate for the damaged area only. The damaged area includes about 9.5 ha. Within this area of the watershed, we estimate a total branch biomass of 55 mg/ha of which 18 mg/ha has been killed. This translates to a 34% reduction in the canopy. We estimate that leaf weight is decreased by 30%. We have not yet made the calculations converting leaf weight into leaf area, but the proportions should not change much from the 30% by weight.

        Using the elemental concentrations in the branches and the estimated killed branch biomass, there are approximately 93 kg/ha of Ca, 23 kg/ha of K and 7 kg/ha of Mg in the dead material. This is 16 %, 13% and 13%, respectively, of the total aboveground amounts of these elements in the biomass (all trees sampled > = 20 cm dbh).

        Not all this material is yet on the ground and not all the possible estimates of amounts have been refined at this time.

        Temperature data are being collected from a network of 20 temperature monitors with four thermistors each (total 80 probes), installed by 26 July 1998. The typical installation has a thermistor at soil depths of 50 cm, 15 cm, and 2 cm, and a shielded air probe at 1-m above land surface. Six of the 20 units substitute a 25-cm soil depth for the air measurements. Temperature monitors were sited next to existing soil water lysimeters, if available, in open gaps, under brush piles, and under undisturbed forest canopy, from lower elevations to ridge-top, in W1 (12 units) and west of W6 (8 units). Litter traps were placed within a few meters of the all Hobo units, and LAI measurements obtained above each site. Temperature loggers have been downloaded at monthly intervals, at the same time that soil lysimeters are sampled. Hourly thermal data are uploaded to a computer spreadsheet, edited for QA/QC and graphed. Thermistor plots were located across the damage zone and from bottom to top of each watershed to record variability due to canopy disturbance and elevation.

        At six of the sites (two of each type), and within 3 m of the temperature monitors, we will measure net N mineralization, nitrification and denitrification rates. Our objective is to measure these processes in undamaged, gap and "brush pile" locations so that we will be able to explore relationships between changes in N cycle processes and leaf area reduction, soil temperature changes and soil solution chemistry in these plots.

        Intact soil cores will be incubated in situ for 30-day periods during spring, summer and fall. Accumulation of inorganic nitrogen during these incubations will provide estimates of mineralization and nitrification and inorganic N pools in the soil. Cores will also be incubated with acetylene and production of nitrous oxide will be measured to provide an estimate of denitrification rates.

        Streamwater samples have been collected for chemical analysis along elevational gradients within the Hubbard Brook Valley, Tenney Mountain area, and upper Pemigewasset River Drainage on a weekly to monthly basis.

        The massive, ice-storm damage to the HBEF provides an opportunity to gain new and valuable information about the ecology and biogeochemistry of nitrogen and carbon, as well as about vegetation dynamics, at several spatial scales. The disturbance "falls on top" of an enormous wealth of data and pre-existing instrumentation for the Hubbard Brook Experimental Forest, which then provides a unique opportunity to evaluate the impact.

        This project is a collaborative effort among C. T. Driscoll and C. E. Johnson of Syracuse University, C. Eagar and C. W. Martin of the USDA Forest Service, D. C. Buso and P. M. Groffman of the Institute of Ecosystem Studies, T. J. Fahey of Cornell University, T. G. Siccama of Yale University; and S. P. Hamburg and A. Rhoads of Brown University. This project is coordinated by G. E. Likens and funding from the NSF is administered through the Institute of Ecosystem Studies.

        We held an HBR-LTER workshop to coordinate this project on 7 April 1998 at the Institute of Ecosystem Studies, followed by an additional meeting at the Annual Hubbard Brook Cooperators' Meeting on 7 July 1998.