11^th Annual Harvard Forest Ecology Symposium - 2000

April 12^th, 2000

We have observed significant year-to-year variability in soil respiration on control plots at the soil warming experiment. Soil temperature can explain much of the variability in soil respiration over the 9 years of the study, but soil moisture appears to limit respiration in dry summers. Relative contributions from the root and microbial respiration to the total soil respiration have remained relatively constant across years irrespective of summer moisture conditions.

Over the 9-year study, annual soil carbon efflux has varied by as much as 30% in consecutive years (Fig. 1). Although soil temperature is generally a good predictor of monthly CO₂ release (Fig. 2a and b), the two years with the lowest CO₂ efflux coincided with years with low summer soil moisture (1995 and 1999) rather than years with low mean annual temperature. Soil temperature measured at 4 cm was able to account for 66% of the variability in monthly CO₂ efflux from the control plots in the soil warming experiment, and 69% of the soil respiration when the years with the two driest summers are excluded (Fig. 2a and b). Soil temperature could only account for 20% of the variability in monthly CO₂ efflux during summers when average soil moisture dropped below 0.15 g H₂O/cm³ soil (1995 and 1999).

The Q₁₀ index, a coefficient describing the exponential relationship between CO₂ efflux and soil temperature, has been variable over the 9 years of the soil warming experiment. The annual Q₁₀ ranged from 1.97 to 3.68 in the control plots. Our measurements suggest that the Q₁₀ index varies predictably with changes in summer precipitation. The cumulative precipitation for the period June through August was linearly correlated with the annual Q₁₀ estimated for the control plots in the soil warming experiment over the last 9 years (Fig. 3).

Although there has been a large amount of variability in annual carbon efflux from the soil, relative contributions of respiration from microbial communities and the root and root-associated rhizosphere have remained relatively constant. A trenching experiment was established in 1994 to quantify the relative contributions of root and microbial respiration to the total carbon efflux from the forest floor. Microbial respiration has accounted for an average of 81% of the total measured soil respiration, and has ranged from 79% in 1998 to 85% in 1996. The dry conditions in 1999 decreased total soil respiration by 20% compared to 1998, but the relative contribution from microbial respiration remained constant (79% in both years).

We are now using these results to improve biogeochemistry models we have developed such as the Terrestrial Ecosystem Model (TEM). We use TEM to study the effects of various kinds of natural and human perturbations of the carbon cycle at regional and global scales.

Losses of carbon in forest ecosystems are likely to be affected by climate and global change. Soil respiration and dissolved organic carbon export are two major pathways for carbon cycling in temperate forests. The effects of chronic carbon and nitrogen manipulation on relationships between dissolved organic carbon export, soil respiration and soil C:N were quantified in the DIRT plot hardwood and Chronic – N hardwood and coniferous stands. Soil solution DOC concentrations were significantly higher in the double above-ground litter input and significantly lower than the zero above-ground litter input (p < 0.01 n = 102). Soil respiration, while significantly lower with zero litter input (p < 0.001 n = 49) was not significantly different from the control plots when double litter was added (p > 0.05 n = 50). With nitrogen additions of 150 kg ha^-1 yr^-1, DOC concentrations in soil solution were significantly higher in both the hardwood and coniferous stands compared to their controls (p < 0.05 n = 204). Soil respiration was significantly affected by high N inputs in the coniferous stand only (p < 0.05 n = 69).

DOC concentration was significantly related to soil respiration in both hardwood and coniferous stands (R² = 0.61 p < 0.0001; R² = 0.92 p < 0.05). Neither manipulation of above-ground carbon inputs nor chronic-N input had an effect upon the relationship between DOC and CO₂ in hardwood forests. When the coniferous stand was fertilized with nitrogen, the relationship between the two variables was lost. Soil C:N ratio was able to explain 75% of the variance in soil solution DOC when hardwood and coniferous stands and treatments were combined (p = 0.01). No relationship was found between soil respiration and soil C:N.

The findings of this study show that manipulation of above ground carbon and chronic nitrogen applications to hardwood forests are unlikely to affect the fundamental mechanisms of DOC production in the short-term. However, there appears to be a de-coupling of the carbon and nitrogen cycles with chronic-N application to coniferous forests. Fertilization of coniferous forests will inevitably lead to reduced surface water quality in terms of increased carbon and nitrogen concentrations.

On a global scale, peatlands are fundamentally important because they contain approximately one third of the earth's carbon pool and play a critical role in global carbon dynamics. In order to understand how peatlands might respond to impending climate change, it is imperative to study how peatland development was influenced by past climatic fluctuations.

With adequate moisture, peatlands form through the mechanisms of terrestrialization and paludification. Terrestrialization, also known as lakefill, occurs when a body of water is slowly filled-in by peat, turning open water into a vegetated community. In contrast, paludification converts upland communities to peatland through lateral extension of the peat mass.

The relative importance of these two mechanisms of peatland development can be distinguished by characterizing the peat deposits and developing detailed chronologies for individual sites. Terrestrialization is identified by the presence of open-water sediments, such as silt or lake mud. The progress of paludification can be conclusively demonstrated by dating basal peat deposits. When a site has paludified, the basal peat deposits are younger towards the edge where expansion has occurred than at the site of initiation.

Both climate and topography function as fundamental controls on the development of peatlands and the dynamics of paludification. It is possible to determine the relative influence of each factor by comparing the ages of basal peat deposits from multiple sites or different areas of the same site. When climate is a driving force across a region, multiple sites may be expected to have similar dates of peat initiation or lateral spread. In contrast, when topography is the dominant influence, the limits of expansion should mirror the basin topography and periods of paludification are asynchronous at different sites.

Worldwide, most peatlands have formed through paludification, rather than lakefill. Yet, based on very general models of wetland formation for temperate regions and due to the high summer temperatures that limit peat accumulation in central New England, terrestrialization has been viewed as the main process of peatland development in the area. The result is that the dynamics of paludification have yet to be examined in temperate areas, despite evidence suggesting that paludification may be fairly widespread and important.

We propose to investigate the developmental history of peatlands in central New England by examining two questions. 1. Has paludification been an important process in the formation of peatlands in central New England? 2. If so, what influence have climate and topography had on the pattern of development and the rates of expansion?

To address these questions, we will study three forested peatlands. At each site, a detailed basin map and descriptions of the sediments will be prepared. Then, a series of basal peat samples will be collected and radiocarbon-dated in order to determine rates of lateral spread.

Micrometeorology-based measurements of forest-atmosphere CO₂ exchange indicate that secondary forest growth in the eastern U.S. currently sequesters about one-third (0.7 Gton yr^-1) of the of atmospheric CO₂ -carbon fixed by forests globally (2 Gton yr^-1). However, this estimate is more than double the forest inventory-based estimate for the same region. The goals of our carbon cycling study are: (1) to interpret the continuous record of net ecosystem exchange of CO₂ (NEE) at Harvard Forest with respect to ground-based measurements of C allocation and efflux; (2) to place our results in the context of past land use, disturbance, and current tree species demography; and (3) to investigate the effects of selective harvest on C cycling in an adjacent forest. Meeting these objectives over several years can assist in predicting forest response to climate variation and in relating NEE to forest history and composition on a regional scale.

Annual NEE in 1998 and 1999 were -1.1 and -2.0 tons C ha^-1, respectively. The timing of ecosystem C uptake and efflux differed greatly, depending on weather features such as early budbreak and late summer drought in ’98, and severe summer drought followed by September rain in ’99. Above-ground wood increment in early successional species constituted a slightly higher proportion of the NEE in ’99 than in ’98 (12% vs. 7%, respectively), possibly due to their predominance in more mesic microsites. However, in both years the above-ground wood increment in all live trees > 10 cm DBH was 60-70% of the whole ecosystem sequestration (0.7 and 1.2 tons C ha^-1, respectively). Oaks (Quercus rubra and Q. velutina) gained well over half the total wood increment in both years.

Within our 40 sample plots (total 1.3 ha), 57 trees containing 4.6 tons C died between 1993 and 1999. Thirty-three of these (3.3 tons C) were red maple (Acer rubrum). In 1999, 23 trees (1.3 tons C) died, including 7 red maples (0.6 tons C). For all tree species in the sample, tree growth into the > 10 cm size class between 1993 and 1999 was far less than mortality in terms of both number of stems and carbon. Our preliminary survey of coarse woody debris (CWD) in the spring of ’99 indicates that the total pool of logs and snags > 7.5 cm diameter at Harvard Forest contained about 6 tons C ha^-1. Our mortality and CWD data together suggest that C storage in dead wood within the fetch of the EMS will increase as a proportion of NEE as the forest matures.

One year of pre-harvest measurements on private property immediately south of the Prospect Hill Tract indicate subtle differences with respect to existing study plots. Total tree biomass is slightly less than in Prospect Hill plots (81 vs. 98 tons C ha^-1, respectively). Biomass distribution among species is slightly different, with hemlock (Tsuga canadensis), birch (Betula spp.) and beech (Fagus grandifolia) comprising a greater proportion of stand biomass. Soil respiration and phenology trends are similar to Prospect Hill.

The hurricane experiment was designed to simulate the impacts of catastrophic wind to mature hardwood forest. In October 1990, canopy trees were pulled over using a winch, resulting in direct and indirect damage to nearly 70% of the stand. One purpose of the experiment is to systematically study the mechanisms of tree regeneration and changes in species composition following such disturbance. Regeneration has been tracked from 1990 (pre-manipulation) - 1999 in 10m² plots along transects in the experimental (0.8 ha) and control (0.6 ha) sites (see Cooper-Ellis et al. 1999 for a complete description of experimental design and methods). We are particularly interested in the role of sprouting in a hardwood-dominated forest, which was an uncommon stand type when the last major hurricane struck this region in 1938.

From 1990 – 1993, saplings and sprouts increased four-fold in the experimental plot (>25,000 stems/hectare), then decreased to 17,500 stems/ha in 1999 (about a three-fold increase from 1990). In the control plot, saplings and sprouts increased 70% from 1990 to 1999; much of this increase was the result of a 1992 winter storm. Red maple, birch species, white ash and black cherry dominate regeneration (Fig. 1). Red oak is less abundant (<5% relative density of saplings and sprouts in 1999) and below average in height. In comparison, the original overstory was composed mainly of red maple and red oak (52% and 19% relative density respectively, with large red oak comprising the majority of the basal area).

Sprouting was an important mechanism for maintaining canopy cover initially, which aided in muting ecosystem-level response to the manipulation (Cooper-Ellis et al. 1999, Foster et al. 1997). However, the long-term importance of sprouting in shaping forest structure and composition is less than advance regeneration and new seedlings/seedling sprouts. As the number of sprouts per tree and number of trees with sprouts have declined, tree sprouts have proportionally decreased, whereas other types of regeneration have proportionally held steady or increased (Table 1).

Ingrowth into the >5cm diameter class gives a preliminary sense of which components of regeneration will successfully establish the new stand. Most ingrowth is advance regeneration, whereas sprouting plays a minor role in forming the new cohort (Fig. 2). Black and yellow birch make up 50% of the ingrowth; red maple and white pine each comprise 10%. We plan to map and measure all new stems >5cm dbh in summer 2000 in the experimental and control plots, so we can track the dynamics of this new cohort as it undergoes self-thinning and interacts with the remaining overstory trees.

Cooper-Ellis, S., D.R. Foster, G. Carlton, and A. Lezberg. 1999. Vegetation response to catastrophic wind: results from an experimental hurricane. Ecology 80:2683-2696.

Foster, D.R., J.D. Aber, J.M. Melillo, R.D. Bowden, and F.A. Bazzaz. 1997. Forest response to disturbance and anthropogenic stress. BioScience 47:437-445.

The Southern Yucatan Peninsular Region contains the most extensive tropical forest in Mexico, but is facing rapid deforestation. Since the late 1960s, government resettlement programs and the construction of a paved highway bisecting the region have led to an influx of people to this region that has been sparsely populated since the collapse of the Mayan civilization 1000 years ago. Harvard Forest is collaborating with Clark University and El Colegio de la Frontera Sur (ECOSUR), Mexico, to document the natural, historical and economic factors driving land use and land cover change in the region. Spatially explicit models of land cover transitions are being developed using remotely sensed imagery from 1975 to the present by researchers at Clark University. However, the satellite imagery begins after major agricultural settlement was well underway. Through the 1900s but especially since 1950, large forestry concessions and forestry ejidos were engaged in selective logging of high-value hardwoods (e.g., mahogany) in the region. As high-value hardwoods have been depleted and agricultural activity has increased, logging has declined. The paved highway was completed in 1967, allowing more agricultural settlements to establish. To extend the history of regional land cover available from remotely sensed imagery, we are interpreting land cover patterns from aerial photographs of the region taken in 1969, capturing spatial patterns at this critical transition in the region’s land-use history.

The black and white, fairly small-scale (~1:80,000) photos cover approximately the northern half of the entire study area (Fig. 1). With the aid of a stereoscope, we delineated all observable roads and apparently disturbed land. The undisturbed forest matrix was not classified into different natural forest types. Land cover was divided into two major categories: agricultural lands (parcel borders sharp and square-edged, and vegetation shorter than surrounding land) or disturbed forest (discontinuous forest cover, which could be a result of human activity or natural disturbance). Each of these major categories was subdivided into finer classifications based on percent vegetation cover. Land-cover and roads are being digitized using ArcView GIS software. Currently, about 80% of the photos have been digitized.

Preliminary observations suggest that agricultural activity was concentrated around older towns on the eastern and western edges of the study region, with little agricultural development elsewhere. Some of the central area with no signs of human use is now part of the Calakmul Biosphere Reserve. The eastern half of the study area (originally the main highway extended only across the eastern half of the region) contains a systematic grid of roads, presumably used for logging access. Many of these roads were faint in the photos and do not appear on current maps of the region.

We plan to relate the broad patterns of land-cover in 1969 to prior logging activity and the more recent patterns of land-cover obtained from the satellite imagery. This data set will provide a more complete understanding of the spatial extent of mid-century logging and the beginning of major deforestation and settlement in this region.

The Montreal Protocol on Substances that Deplete the Ozone Layer in 1987 and its subsequent London (1990) and Copenhagen (1992) Amendments mandated control measures on the production and consumption of ozone-depleting substances [UNEP, 1985-1997]. The majority of the substances, including CFC11, CFC12, CFC113, Halon-1211, CCl₄, and CH₃CCl₃, were scheduled for 100 percent reductions in production and sales by 1 January 1996 in developed countries. The success of the Protocol in the U.S. has thus far been determined by inventory estimates only. Given that today emissions violate international agreements, they may not be reported willingly. The quality of these inventories and the recent urban pollution history has yet to be independently established for the post-1996 ban years.

To address this deficiency, this study provides an independent measure of emissions from a major emitting region of ozone-depleting species for the years 1996 through 1998, the first three years after the full implementation of the Montreal Protocol. The measurements were taken every 24 minutes at Harvard Forest, MA, downwind of the Northeast urban-industrial corridor, including the greater metropolitan region of New York City. Using the well-documented EPA carbon monoxide emissions, which are reported on a per county basis, and a composite PCE inventory, derived from the EPA/TRI records and the McCulloch and Midgley sales-based country-level tallies [McCulloch et al, 1996 and P. Midgley, personal communication], we estimate the annual and seasonal urban/industrial emissions of CFC11 (CCl₃F), CFC12 (CCl₂F₂), CFC113 (C₂Cl₃F₃), methyl chloroform (CH₃CCl₃), chloroform (CHCl₃), carbon tetrachloride (CCl₄), PCE (C₂Cl₄), and halon-1211 (CBrClF₂), as well as hydrogen (H₂), CO, methane (CH₄), nitrous oxide (N₂O), and sulfur hexafluoride (SF₆), all on a per capita basis. The results of this study confirm the accuracy of the above listed inventories for the New York City – Washington, D.C. corridor.

The seasonal character of the urban/industrial emissions for the New York – Washington, D.C. corridor are calculated as functions of CO and PCE releases. Seven cases are considered: all data; northwest winds only (NW); southwest winds only (SW); day (6 am to 6 pm); night (6 pm to 6 am) ; high U* (>0.2 m/s); and low U* (<0.2 m/s) (where U* = (Ö momentum flux)/100, a measure of how well mixed the air is). The two cases of ‘northwest winds only’ and ‘southwest winds only’ are chosen because the Harvard Forest station receives most of its air flow from the west, with the most polluted air from the southwest and the fastest winds and least polluted (background) air from the northwest. All seven cases track each other well. Only the NW case shows a tendency for non-conformity, particularly for H1211, CHCl₃, and N₂O. Seasonal cycles exist in the emissions of CFC12, CFC11, and H1211, with highest values consistently found in the summer and lowest in the winter. In the case of CFC12, such a cycle is to be expected given that its primary use is as a coolant in car air-conditioners which are used in the summertime. The pollution of SF₆ also appears to have a seasonal cycle with highs in the spring and lows in the summer/fall. The high degree of similarity between CO/PCE and H₂/PCE suggests that CO and H₂ are emitted by the same primary and secondary sources. The same may be said about CFC11 and CFC12, whose ratios to PCE are strikingly similar.

The annual urban pollution emissions for each species indicate that a number of the species exhibit distinct inter-annual trends. Carbon monoxide, CFC11, N₂O, and TCE appear on the whole to be increasing. For CFC11, such a positive trend is a surprise, especially given the ban on its production by the Montreal Protocol, the decline in its atmospheric growth rate witnessed as early as 1989 [Elkins et al., 1993], and its close seasonal correspondence to CFC12 whose emissions are decreasing, not increasing. Hydrogen’s urban/industrial pollution emissions appear stable, and Halon-1211 and CHCl₃ do not exhibit any distinguishable pattern. Methane, CH₃CCl₃, PCE (in all categories except 1998 CO-all year), and SF₆ all decrease over the years 1996, 1997, and 1998.

The decline in methyl chloroform, although particularly dramatic, is tailing off slower than was estimated by inventories (Figures A and B) [Bakwin, 1997; Derwent, 1998]. This is a noteworthy feature that will benefit modellers concerned with seasonal cycles for OH computations. A linear extrapolation from the 1994 and 1995 North America sales-derived values along with the 1996, 1997, and 1998 FACTS-based data implies that CH₃CCl₃ emissions ended in early 1999. An exponential curve through all points of the North America precipitous drop (1990-1995) and of FACTS predicts that CH₃CCl₃ emissions will not cease until as late as the year 2004. If such is the case, the CH₃CCl₃ lifetime of 4.8 years suggests that the end of the absolute CH₃CCl₃ concentrations may be anticipated by 2010, with the atmospheric OH abundance responding accordingly.

Bakwin, P. S., D. F. Hurst, P. P. Tans, and J. W. Elkins. 1997. Anthropogenic sources of halocarbons, sulfur hexafluoride, carbon monoxide, and methane in the southeastern United States. J. Geophys. Res. 102: 15915-15925.

Derwent, R. G., P. G. Simmonds, S. O’Doherty, and D. B. Ryall. 1998. The Impact of the Montreal Protocol on concentrations in northern hemisphere baseline and European air masses at Mace Head, Ireland over a ten year period from 1987-1996. Atmospheric Environment 32: 3689-3702.

Elkins, J. W., T. M Thompson, T. H. Swanson, J. H. Butler, B. D. Hall, S. O. Cummings, D. A. Fisher, and A. G. Raffo. 1993. Decrease in the growth rates of atmospheric chlorofluorocarbons 11 and 12. Nature 364: 780-783.

McCulloch, A. and P. M. Midgley. 1996. The production and global distribution of emissions of trichloroethene, tetrachloroethene and dichloromethane over the period 1988-1992, Atmospheric Environment 30: 601-608.

United Nations Environmental Programme. 1985. Vienna Convention for the Protection of the Ozone Layer, 1987 Montreal Protocol to Reduce Substances that Deplete the Ozone layer Report, Final Report (New York), 1990 London Amendment, 1992 Copenhagen Amendment, 1997 Production and Consumption of Ozone Depleting Substances, 1986-1995, Nairobi, Kenya.

Photosynthesis plays a key role in the forest carbon cycle because it represents the primary process of terrestrial productivity. At the leaf level there is a strong connection between maximum photosynthesis and nitrogen availability because nitrogen is needed for the synthesis of Rubisco and chlorophyll. This fundamental relationship has been intensively used to scale maximum rates of photosynthesis from leaf level to stand and biome level. However the general use of this photosynthesis-nitrogen relationship has widely ignored the fact that not all of the nitrogen present in a leaf or a needle is "functional", i.e., is directly involved in the primary processes of photosynthesis. At the whole plant level nitrogen is also needed to promote growth and nutrient uptake. In this way several physiological processes compete with photosynthesis for available N. Therefore, if we want to scale photosynthesis on the basis of plant nitrogen concentration, we need to know how plant internal partitioning of N is regulated and how changing partitioning patterns affect leaf photosynthesis.

We have initiated a study in which we investigate the effects of changes in nitrogen partitioning on the photosynthesis-nitrogen relationship of forest trees. By following the seasonal dynamic of leaf photosynthesis and N partitioning we will be able to investigate whether physiologically active pools of N in the foliage can be used to scale leaf photosynthesis. Based on earlier studies carried out at the Harvard Forest we hypothesize that the photosynthesis-nitrogen relationship will change over the course of a growing season. Preliminary data from two of our study sites at Harvard Forest and Howland Forest (ME) confirmed these expectations. Individual species such as hemlock and red spruce showed a different slope in the photosynthesis-nitrogen relationship than, e.g., red oak and red maple. A comparison of protein and Rubisco concentrations for the study species supports our initial hypothesis that e.g. nitrogen investment into Rubisco is not constant across species and probably throughout the growing season. Therefore the seasonal change in photosynthetic nitrogen use efficiency has to be taken into account when scaling leaf photosynthesis to whole canopies.

Rich mesic forests are uncommon upland communities that support numerous rare plant species and have been identified by the MNHESP as high priorities for conservation in Massachusetts (Swain 1999). However, the factors controlling the spatial distribution and species composition of this community are poorly understood. Rich mesic forest (RMF) sites are characterized by uncommon environmental conditions, especially mesic, circumneutral soils (Weatherbee 1996), that support a unique and diverse herbaceous flora including many species that may be particularly sensitive to the impacts of past land-use (Henry et al. 1974, Bormann et al. 1979). The disturbance-sensitive nature of the herbaceous flora suggests that RMF may have occupied a more extensive portion of the landscape prior to widespread forest clearance for agriculture during the 18^th and 19^th C (Jenkins 1994). The RMF community is considered to be a northern variant of the Mixed Mesophytic Forest Type of the southern Appalachian Mountains, which has also been noted for its diverse herbaceous layer (Braun 1950). In Massachusetts, RMF sites are most common in the western portion of the state, where their distribution largely coincides with the occurrence of calcareous bedrock.

Evaluating the contrasting roles of historical human disturbance and environmental conditions on the spatial distribution and species composition of the RMF community is a critical aspect of understanding the causes of this community’s limited distribution in the state and the rarity of many plant species associated with it. To address these issues, a broad scale survey of extant RMF sites in Franklin and Hampshire Counties, Massachusetts will be conducted to quantify environmental factors controlling the community’s distribution and plant species composition. Using environmental data gathered in the broad scale survey and land-use history records from the 19^th and early 20^th C, a GIS model will be developed to investigate the influence of past human disturbance on the modern distribution of RMF. The model will be developed and tested for two towns in western Franklin County that are known to support numerous occurrences of the RMF community. Such an approach will lead to a characterization of RMF distribution and composition in the region, detailed information about numerous sites in western Franklin and Hampshire Counties, and increased understanding of the factors controlling species distributions and community patterns. As a result, this study will provide information that is critical for the long-term conservation and management of the RMF community.

Bormann, F.H. and Likens, G.E. 1979. Catastrophic disturbance and steady state in northern hardwood forests. American Scientist 67: 660-669.

Braun, L. 1950. Deciduous forests of eastern North America. The Blakiston Company, Philadelphia.

Henry, J.D. and Swan, J.M. 1974. Reconstructing forest history from live and dead plant material- an approach to the study of forest succession in southwest New Hampshire. Ecology 55: 772-783.

Jenkins, J.C. 1994. Conference notes from a symposium on rich woods. Unpublished.

Swain, P. 1999. Draft of Terrestrial and Estuarine Plant Community Classifications for Massachusetts. Massachusetts Natural Heritage and Endangered Species Program. Unpublished.

Weatherbee, P. 1996. Flora of Berkshire County, Massachusetts. The Berkshire Museum, Pittsfield, MA.

There is growing evidence to suggest that the impact of N deposition on temperate forest productivity, even in the absence of N saturation, may not be as great as several studies suggest. Further, most studies of the impact of N deposition on forest processes have ignored changes that take place in the canopy. To date, there have been no empirically-based, ecosystem-scale studies of how N deposition alters canopy processes and thus the potential for C gain in forests. We initiated a study of how N deposition affects the physiology, dynamics, and structure of temperate forest canopies. We are examining how the partitioning of N within tree foliage, how it is affected by N deposition, and how these changes in N allocation relate to changes in photosynthetic capacity. In addition we are examining the effects of N deposition on foliage retention time and canopy structure in order to begin scaling changes in leaf-level processes to the entire canopy. Our study is based at the Harvard Forest Chronic N Experiment, where a long-term (12-yr+) of N deposition (+5 and +15 g N m^-2y^-1) on both coniferous and deciduous forests is underway. Our preliminary measurements (within a Pinus resinosa stand) demonstrate that foliar N content has significantly increased, and that this increase is accompanied by a de-coupling of the Photosynthesis-N relationship. Preliminary evidence suggests that the de-coupling is due either to increased production of non-photosynthetic proteins (e.g., Calvin cycle enzymes) or an excess production of Rubisco. The net effect of this de-coupling is that the large increases in foliar N we have observed are not accompanied with increased photosynthetic capacity. Further, we have found that needle life-span has been significantly reduced, and, even with apparent increases in foliar production (at the expense of wood production), standing leaf area has decreased. Taken together, these data suggest that potential C gain by the Pinus resinosa canopy has been significantly reduced by long-term N deposition.

Over the past year, the Harvard Forest web page (http://lternet.edu/hfr) was redesigned and a new system implemented for the organization and presentation of research data on-line. The main features of this system (which is still under development) are outlined below.

An electronic Data Archive is maintained at the Harvard Forest for all scientific projects since 1988 (beginning of LTER I), as well as selected earlier projects. The Data Archive is stored as a web-compatible file system and can be viewed with a web browser such as Netscape. A copy of the Data Archive is maintained on-line as part of the Harvard Forest web page.

The Data Archive can be accessed and searched via the Harvard Forest Data Catalog (http://lternet.edu/hfr/data/catalog.html), which appears on the Data Archive homepage. The Data Catalog contains a list of research projects arranged by research category. Each project entry includes project title, list of investigators, and project code. Each entry is linked to a separate project page, which contains additional metadata (project dates, field site, and contact person), a project Overview, and (for completed projects) links to Data and Metadata files. All files in the Data Archive are stored in plain ASCII or HTML format (except for spatial data, which are stored in ArcView or Idrisi format). A few projects are maintained on web servers at allied institutions, with links from the Harvard Forest web page. Permanent copies of the Harvard Forest web page, including the entire Data Archive, are created annually and stored at the Harvard Forest and an offsite location.

Selected metadata for each project, including project title, investigators, keywords, and project code, can be searched on-line via the LTER Network Data Catalog (http://lternet.edu/DTOC). Harvard Forest metadata are stored in a separate location (http://lternet.edu/hfr/data/dtoc.html) for weekly harvesting by the Network Data Catalog.

A Projects database is also maintained offline for Data Management purposes. For each project, this database contains the metadata described above (project code, title, list of investigators, contact person, keywords, dates, and field site) as well as project status and access codes for relevant materials in the Document and Sample Archives. The Projects database is used to track the progress of current research projects and to ensure that files in the Data Archive are updated in a timely manner.

For the past year, archive responsibilities have been divided among the following areas: archive and records management, historical research, and administrative projects. The following is a summary of the major archive related projects.

To continue the effort to make backups of crucial Harvard Forest historical records, a large microfilm project was recently completed. The material which was duplicated included all forest inventory records from 1907 through 1992, 1938 Hurricane historical information and salvage data, plantation, nursery, Cabot Foundation and case history files, outlot records for the Pisgah, Matthews Plantation, Schwartz Tract, and Tall Timbers properties, and cutting summaries. The microfilm rolls, along with those for the stand records and 1830s map series, are stored in the Shaler Hall vault, with duplicates at the Harvard University Archives in Cambridge.

In conjunction with the renovation of the Shaler Hall basement, all of the historical reprint collections were collated, sorted and inventoried. A large library of Harvard Forest reprints from the 1900s to the present was established in the basement in archival folders and new cabinets, arranged in chronological order to facilitate access by researchers. During the sorting and purging process, a considerable amount of material was relocated to appropriate individuals and institutions, including Harvard Forest and collaborative researchers, the Arnold Arboretum, the University of Southern Maine Map Library, the Arctic Institute at the University of Edmonton, the Raup and Zimmermann families, and Ann Lewis at the University of Massachusetts Department of Natural Resources.

New material continued to strengthen the Archive holdings. Over 150 towns from the Massachusetts 1830s series were added to the map collection, and an updated and reformatted map catalog was printed. The extensive physiology research collection, including film logs, photographs, figures, and slides, of Dr. Martin Zimmermann was incorporated into the archive research file collection. Important historical photographs and slides were obtained from the Massachusetts DEM office in Clinton, David Tatlock, Larry Buell, and Albert Upham. Additions were made to the research, publication, and sample archive collections on a regular basis.

The Archives continued to be an important element of the Harvard Forest, as the material and facilities were used on a daily basis by staff, collaborators, local residents, and local organizations such as the Metropolitan District Commission, Massachusetts Department of Environmental Management, Worcester EcoTarium, and North Quabbin Greenway planners.

The surveyor notes in early New England city and town records offer a valuable source of information about presettlement forest composition at the township level. The most comprehensive records are found in early land abstracts, usually known as proprietor books; in some towns tree species citations have totalled 2000 or more. When early land data is unavailable, information has been collected from highway and boundary surveys from town record compilations. In instances where no data was found for a particular town, this indicates that either the early records were lost or destroyed, or surveyors did not cite individual tree species in their lot or road surveys.

Prior to 1999, New England witness tree data had been collected for Vermont, north-central Massachusetts, the northern and southern Berkshires, Cape Cod, and parts of New Hampshire. In the past year, additional data has been gathered from New Hampshire and all regions of Massachusetts not covered in past studies. In Massachusetts, J. Burk and E. Largay have compiled data from several sources, including records at the Massachusetts Archives, county registries of deeds, and individual town halls and libraries. Detailed proprietor data was found for the Connecticut River Valley towns, much of Hampden County, southern Worcester County, and the eastern communities. Information from central Berkshire and western Franklin and Hampshire Counties was not as comprehensive and was mainly compiled from road surveys. The biggest gap in information was eastern Berkshire County. At the New Hampshire State Library, J. Burk gathered proprietor data for 20 towns in the south-central region, with the most detailed records found in the Concord-Manchester area and along the Connecticut River. This information was used by C. Cogbill as part of a manuscript for the Rhodora Symposium.

Once all remaining data for Eastern Massachusetts is compiled, the next states to be addressed will be Connecticut and Rhode Island. All information will be sorted by state and county and archived at the Harvard Forest as a resource for use in present and future studies.

Despite increasing levels of nitrogen deposition at temperate latitudes, no studies to date have considered how this perturbation will alter temperate forest species composition. Nitrogen deposition may affect mixed forests in particular due to differences in nutrient use patterns between evergreen coniferous and deciduous broad-leaved species. To examine how nitrogen deposition may alter the composition of mixed conifer broad-leaved forests, we investigated regeneration responses of both types of species to increasing nitrogen, and additionally compared these responses across coniferous (hemlock) and broad-leaved (red oak) stand types. We addressed nitrogen effects on two critical components of forest regeneration dynamics: (1) development of an understory seedling bank, and (2) seedling responses to canopy gap formation. We applied nitrogen (0, 2.5, 7.5 g m^-2 yr^-1) both to replicated understory plots in three hemlock and three red oak dominated stands, and to seedlings growing in high light conditions in soil from each stand. In both cases, we examined seedling survival and growth for three coniferous and three broad-leaved species over two years.

Nitrogen differentially influenced species’ regeneration patterns, although the nature of these effects depended on light environment. In the understory, changes in seedling bank structure were driven primarily by differential seedling emergence and survival, with nitrogen decreasing red maple and red spruce abundance, and increasing birch relative abundance (Fig. 1). These changes were particularly significant in hemlock dominated stands. Under gap conditions, nitrogen predominantly increased seedling growth, with early-successional broad-leaved species responding most strongly, and coniferous species remaining relatively unresponsive (Fig. 2). In this case, seedlings were generally more responsive to nitrogen on soil from red oak stands. Thus, increasing nitrogen deposition will alter the composition of mixed temperate forests. However, given that seedling responses to nitrogen were contingent on prevailing light and soil conditions, the exact nature of the change will depend on both current forest stand composition and future patterns of disturbance.

We investigated how light and CO₂ levels interact to influence growth, phenology, and the physiological processes involved in leaf senescence in red oak (Quercus rubra) seedlings. We grew plants in high and low light and in elevated and ambient CO₂. At the end of three years of growth, shade plants showed greater biomass enhancement under elevated CO₂ than sun plants. We attribute this difference to an increase in leaf area ratio (LAR) in shade plants relative to sun plants as well as to an ontogenetic effect as plants increased in size, the LAR declined concomitantly with a decline in biomass enhancement under elevated CO₂ (Fig. 1). Elevated CO₂ prolonged the carbon gain capacity of shade grown plants during autumnal senescence, thus increasing their functional leaf lifespan (Fig. 3). Prolongation of carbon assimilation, however, did not account for the increased growth enhancement in shade plants under elevated CO₂. Elevated CO2 did not significantly alter leaf phenology (Fig. 2). Nitrogen concentration of both green and senesced leaves was lower under elevated CO₂ and declined more rapidly in sun leaves than in shade leaves (Fig. 4B). Similar to nitrogen concentration, the initial slope of A/C_i curves indicated that Rubisco activity declined more rapidly in sun plants than in shade plants, particularly under elevated CO₂ (Fig. 5). Absolute levels of chlorophyll were affected by the interaction of CO₂ and light, and chlorophyll content declined to a minimal level in sun plants sooner than in shade plants (Fig. 4A). These declines in nitrogen concentration, in the initial slope of A/C_i curves, and in chlorophyll content were consistent with declining photosynthesis, such that elevated CO₂ accelerated senescence in sun plants and prolonged leaf function in shade plants. These results have implications for the carbon economy of seedlings and the regeneration of red oak under global change conditions.

Feedbacks between biotic processes and climate are one of the largest unknowns in the climate change debate. One of the direct consequences of global climate warming is alteration of energy balances of ecosystems, which in turn can produce significant effects on physical and ecological processes. Understanding underline mechanisms and specific feedbacks of each energy term, hence, is vital in explaining overall responses of an ecosystem (e.g., productivity and stability). A full-factorial design was constructed in 54 mesocosm plots by manipulating heat loading (3 levels) and water table (3 levels) of two dominant peatlands in southern boreal forest with three replications. Our central hypothesis is that climate forcing of heat loading and water-table depth determine plant community and ecosystem structure in northern peatlands, which in turn have a feedback effect on the thermal and radiative energy budgets of the system.

Energy fluxes of net radiation (Rn) and soil heat (G) and vertical soil temperatures (Ts) at 5, 10, 25, and 40 cm in all 54 plots were continuously measured every 20 second and their averages stored at 30-min intervals since October, 1998 using an automated energy budget system (Fig. 1) to detect underlying mechanisms causing temperature changes. Cover of vascular plants, and cover and vertical growth of Sphagnum were measured during the summer of 1998, and weekly snow depths were measured throughout the snow season. Data from January 1 through December 31, 1999 are included in this report.

We found no significant difference in Rn (P<0.02) within any treatment but among treatments. Large differences in G and Ts were recorded between fens and bogs at several temporal scales from hours, days, and seasons, suggesting energy balance of two ecosystems have distinctive vegetation-energy feedback. This conclusion is also supported by significant linear correlation between energy terms (Rn & G) and vegetation characteristics (plants, thatch, snow, and Sphagnum). In general, fens were warmer than the bogs at all depths and they also thaw earlier in spring.

Bogs appeared to be more sensitive than fens in Rn, G, and Ts. One important founding of this experiment is that heat loading produced the greatest effect during the winter and spring, with smaller difference in the fall. During the winter, soils of heated plots were often cooler than unheated plots, likely caused by reductions of snow cover. This is especially true with high water table in fens. In the spring, we detected the highest temperature differences between heated and unheated plots, with very strong water table interactions (Fig. 2). Ts at 40 cm in heated plots were >0 ºC in early March, 20 to 40 days before unheated plots reached the same level in April. Finally, heat loading elongated the growing season by 10-30 days (i.e., 20-25% increment) depending on the loading level and water table.

In summary, our results suggested that global warming can cause soil cooling during the winter. Such unexpected effects will likely happen in the southern-edge of snow zones in temperate areas, or in years when snow fall is significantly less (or more). Results from 1998 have triggered many critical questions and research topics. For example, if the length of growing season is increased, would ecosystem productivity be increased at the same level? Obviously, our future challenge lies in linking the biophysical changes to ecological processes such as canopy dynamics, decomposition, and gas exchange in hope that a responsive model under a changed climate can be developed.

Experiments were conducted to compare the affinity and reactivity of three different minerals for natural organic matter (NOM) in forest floor leachate (FFL) from hardwood and pine stands at the Harvard Forest, Petersham MA (Chorover and Amistadi, 2000). The FFLs were acidic (pH 4) with ionic strengths of 1.4 mM (hardwood) and 1.1 mM (pine), and they contained larger organic molecules (weight average molecular weights [Mw] = 5-6 kDa) than has been reported recently for surface waters using similar methods. A synthetic diluent solution was prepared to match the inorganic chemistry of the FFL and to provide a range of initial dissolved organic carbon (DOC) concentrations (0 to 140 g C m^-3) for reaction with goethite (a-FeOOH), birnessite (d-MnO₂) and smectite (montmorillonite, SWy-2) in suspension, and in corresponding blanks.

A variety of macroscopic and spectroscopic methods were employed to show that reaction with the three minerals resulted in distinctly different NOM adsorption, fractionation and transformation patterns. Goethite exhibited a steep initial slope in the adsorption isotherm and a maximum retention of 10.5 g C kg^-1. The isotherm for montmorillonite was more linear, but equal amounts of C were adsorbed to goethite and montmorillonite (per unit sorbent mass) at maximum DOC (Fig. 1). Whereas preferential uptake of high Mw , aromatic constituents via ligand exchange was observed for goethite, compounds of lower than average Mw were retained on montmorillonite and no preference for aromatic moieties was observed (Figs. 2-3). In addition, infrared spectroscopic (FTIR) data indicate that montmorillonite exhibits a higher affinity than goethite for aliphatic constituents of NOM. Birnessite, which has an isoelectric point of pH < 2, retained low amounts of organic C (< 2 g C kg^-1) but showed the highest propensity for oxidative transformation of the NOM (Fig. 4). Oxidation of NOM was coupled to reduction and solubilization of Mn(IV) and Fe(III).

The data indicate that, upon contacting mineral soils, dissolved NOM is fractionated into adsorbed and soluble forms in a manner that is highly dependent on mineral surface chemistry. The affinity of a given clay for organic C does not itself provide a satisfactory measure of which NOM constituents are adsorbed versus which remain in solution. Therefore, mineralogy can strongly impact (i) the types of organic molecules that are sequestered in subsurface mineral soils and (ii) the organic chemistry of drainage waters. This work also emphasizes the fact that abiotic redox transformation reactions must be considered in studies of NOM interaction with Fe(III) and Mn(IV) containing solid phases. These reactions contribute to the oxidative degradation of NOM and result in the transient accumulation of kinetically-active reducing power (e.g., Fe(II) and Mn(II)) in the soil solution, even under oxic conditions.

Chorover, J. and M. K. Amistadi. 2000. Reaction of forest floor organic matter at goethite, birnessite and smectite surfaces. Submitted to Geochim. Cosmochim. Acta.

Paleoecological evidence indicates that the forests of New England were changing in composition before European arrival. In particular, several pollen records from lakes in the Northeast show declines in beech and hemlock coincident with a period documented in historical and dendroecological records known as the Little Ice Age (~1450-1850). This interval is characterized by highly variable conditions, with greater frequencies of long cold winters, and short cool summers. Cooler temperatures alone are not consistent with declines in beech and hemlock, both northern temperate tree taxa. Other related factors, including moisture, may contribute to these species decline. In order to comprehensively investigate this pre-settlement vegetation trend, we are analyzing the paleo record of multiple proxy indicators of climate from several sites in New England, spanning both latitudinal and topographical gradients. Fine-resolution pollen and charcoal analyses allow us to reconstruct vegetation and fire histories, while independent climate proxies - fossil diatoms, chironomids, and stable isotopes - are indicative of aquatic and hydrological conditions.

Levi Pond in Groton, Vermont was selected as the northernmost study site. The land outside of the village of Groton, settled in the late 1700’s, was considered too rocky to farm and the major cultural impacts on this landscape have been logging, and the railroad. Today Levi Pond is surrounded predominantly by red spruce, hemlock, white pine, white birch, and red and sugar maple.

Pollen analysis of the top one meter of sediment from Levi Pond representing the last 3000 years, indicates a gradual increase in spruce to the present (Fig. 1). Beech and hemlock decline coincident with settlement, while birch increases. The diatom record shows changes coincident with the long-term increase in spruce, settlement, logging activity, and more recent development of camps around the site. The fossilized remains of chironomid insects reflect lake-water surface temperatures. In combination with information gained from stable isotope analysis on the lake and catchment hydrological balance, these indicators will provide independent climate assessments in relation to the Little Ice Age. At this site, the pollen record shows declines in beech and hemlock that follow the Little Ice Age period.

The hemlock woolly adelgid (HWA), an aphid like forest pest introduced from Asia, is spreading virtually unrestricted through the eastern hemlock (Tsuga canadensis) range in northeastern North America. T. canadensis does not appear to have any resistance to HWA, nor has an effective natural HWA predator been identified in this region. This epidemic has the potential to dramatically alter the ecosystem function of T. canadensis forests and may influence nutrient cycling in subsequent stands. A recent study by Jenkins et al. (1999) suggested that changes in decomposition rates may be partially responsible for alterations of ecosystem function in HWA infested stands. Decomposition of foliar litter plays a critical role in the assimilation and release of elements through forest ecosystems. Decomposition rates are closely associated with foliar quality and a growing body of knowledge suggests that insect attack can alter foliar chemistry. Understanding decomposition in HWA attacked ecosystems will be critical for determining how this infestation alters the function of T. canadensis ecosystems.

The objectives of this study are: (i) to determine if HWA attack influences decomposition by directly altering foliar quality, and (ii) to determine if HWA attack influences decomposition by altering microclimate. To accomplish this we designed an 18-month in situ decomposition study. We placed two sets of mesh bags containing T. canadensis foliage on the soil surface at each of eight study sites (see Orwig et al., this volume). One set of bags contained foliage from yet uninfested stand at Harvard Forest while the second contained foliage from the individual study site.

Six-month results indicate that HWA attack is slowing decomposition in stands with severe and moderate infestation levels. Although the trend was more pronounced in the severely infested sites compared to the moderately infested sites, decomposition was indistinguishable between the two types of litter. This suggests that microclimate is having a much greater influence on decomposition than initial foliar quality. Although soil temperatures are similar between all sites, the severely and moderately attacked stands have lower soil moisture. These results suggests that dry soil is limiting the establishment of fungal hyphae and microbial communities on soil surface foliage and is thus slowing decomposition in these sites.

Jenkins, J.C., J. D. Aber, and C. D. Canham. 1999. Hemlock woolly adelgid impacts on community structure and N cycling rates in eastern hemlock forests. Can. J. For. Res. 29: 630-645.

Since late 1991, we have made more than 190,000 discrete analyses of ambient methane (CH₄) concentrations. Our automatic gas chromatographic system samples every 6-15 minutes (every 15 minutes since 1996) and samples are compared to working standards calibrated against NOAA-CMDL standards. The observation mixing ratios ranged from 1.709 to 2.471 ppmv, a difference of 762 ppbv. The daily averages varied from 1.765 to 2.045 ppmv (Fig. 1), with monthly medians varying over a much narrower range, 1.79- 1.88 ppmv.

Seasonal trends in the monthly medians are evident in each of the years that we sampled (Fig. 2). The minima occurred in spring and summer and the maxima in fall/winter. Minima appear to be due to increases in photochemically produced tropospheric oxidants before biogenic sources warm enough to produce sufficient CH₄ to significantly affect ambient mixing ratios. Significant variability in seasonal amplitude and timing is present from year to year.

Long term trends in the data were masked by the occurrence of one to three day events of higher mixing ratios when ambient CH₄ would increase by more than 200 ppbv over background levels. Variability in ambient CH₄ mixing ratios is correlated with wind direction, with highest average mixing ratios occurring during time periods when air reaching the tower is from the southerly and southwesterly directions (Shipham et al., 1998b). Examination of the lowest 10% quantile of the CH₄ data reveals a long term increasing trend in these "background" data (Fig 3) of 4.3 ppbv per year in the monthly medians from November 1991 to December 1997, with an r² of 0.35, in general agreement with observed global clean air trends. Similar calculations using the upper 10% of the data suggest that there is little long term increase (r² = 0.0002) and that there are no strong trends in the most polluted air reaching the tower.

The sporadic high CH₄ events at the tower occur more frequently in the winter and thus have a striking effect upon the amplitude and the shape of the seasonal CH₄ signal. Seasonal trends in median values are more clearly revealed through examination of trends in the highest and lowest 10% quantiles of the data, which differ both in magnitude and timing (Fig. 4). Median values for the lowest 10% exhibit a seasonal amplitude of roughly 33 ppbv with maxima in late winter, February and March. Monthly medians of the top 10% exhibit a seasonal amplitude of 113 ppbv.

A diurnal signal is also evident in the data after the inter-annual and seasonal signals are removed (Fig 5). Diurnal maxima occur at night and minima during the day differing by about 12 ppbv. Again there appear to be seasonal differences in both the amount of diurnal variation and the times of daily maxima (Shipham et al., 1998a). Mixing ratio maxima occur near or just after sunrise in the spring and summer, but much closer to sunset in the fall and winter months.

Shipham, M.C., K.B. Bartlett, P.M. Crill, R.C. Harriss and D. Blaha. 1998a. Atmospheric methane measurements in central New England: An analysis of the long term trend and the seasonal and diurnal cycles. J. Geophys. Res. 103: 10621-10630.

Shipham, M.C., P.M. Crill, K.B. Bartlett, A.H. Goldstein, P.M. Czepiel, R.C. Harriss and D. Blaha 1998b. Atmospheric methane measurements in central New England: An assessment of regional transport from surrounding sources. J. Geophys. Res. 103: 21985-22000.

D. Dail, K. Dria, E.A. Davidson, J. Chorover and P. Hatcher

Biotic and abiotic mechanisms by which inputs of inorganic-N are converted to organic-N and retained in soil were investigated using ¹⁵N in sterilized and live soils. Soils were obtained from areas outside the chronic N deposition plots at Harvard Forest. Samples of the control and sterilized organic soil were analyzed immediately after ¹⁵N addition and 24 hours later to determine ¹⁵N incorporation of nitrate or nitrite into inorganic and organic soil-N pools.

Up to 80% of ¹⁵N-nitrite was not recoverable immediately after addition to sterilized organic soils, while 10-20% loss was observed in control soils (Fig. 1). About 30-50% of the nitrite was recovered in an organic fraction during the 24 hour incubation in both live and irradiated soils, indicating an abiotic process, such as nitrosation of soil organic matter, may be responsible for this immobilization (Azhar et al., 1986). Although sterilization by autoclaving or g -irradiation did not significantly affect the relative percentages of label recovered in inorganic and organic pools, autoclaving reduced overall recovery of ¹⁵N immediately upon addition. We suspect conversion to gaseous N may account for the N loss observed when nitrite was added to both control and sterilized soils.

Added ¹⁵N-nitrate was recovered largely as-is (>60%) in all soil treatments and less than 8% was recovered as organic-N with no significant changes after 24 h incubation (Fig. 2). Less than 1% of ¹⁵N-nitrate was recovered as ammonium and we have not included this pool in the figures for sake of simplicity. Similar amounts of ¹⁵N were recovered in the organic pool in sterilized and control soils, however organic N retained in the soil after mineral-N extraction and mineral nitrogen in the extract did not sum to the total ¹⁵N added to the sample (Fig. 2). We report a substantial fraction of nitrate (30 to 50%) was not recovered as inorganic-N or insoluble organic-N. The deficit N, we believe, was removed from the soil by the mineral-N extraction but was not in the form of nitrate, nitrite or ammonium. We have termed this fraction "soluble-N" and suspect it may be organic, but have not made a direct determination of the form of N in this pool.

Substantial ¹⁵N added as nitrate was calculated (by difference) to be in the soluble-N pool in both control and sterilized soil, with no differences observed over the course of the incubation period. This suggested that the nitrate immobilization process into this pool might be abiotic. Sterilization of soil may lead to changes in soil properties that enhance the conversion of nitrite and nitrate to nitrogen gasses or unknown N forms. Hence, artifacts of sterilization are a concern, but nevertheless, these results show that an unidentified process of abiotic immobilization of nitrate to form organically bound N may be important in organic layers of temperate forest soils. These findings support observations by Berntson and Aber² that show nitrate immobilization, as measured by disappearance of ¹⁵N-nitrate from soil cores, was best described by 2 kinetic rates. An immediate, (<1 h) rapid rate accounting for up to 80% of nitrate immobilization and a much slower rate approximating what has traditionally been attributed to microbial nitrate immobilization.

E.S. Azahar, R. Verhie, M. Proot, P. Sandra & W. Verstraete (1986). Binding of nitrite-N on polyphenols during nitrification. Plant and Soil 94:369-382.

G. M. Berntson and J.D. Aber (2000). Fast nitrate immobilization in N saturated temperate forest soils. Soil Biology & Biochemistry 32:151-156

Samples from canopy-to-soil stratigraphic profiles, containing leaf, fine root and organic and mineral soil samples, were collected from the Harvard Forest Chronic N plots during 1989 to 1993 by Knute Nadelhoffer and coworkers. These samples were analyzed for chemical structural information by solid-state ¹³C and ¹⁵N NMR and molecular level detail by pyrolysis GC/MS. Results were used as a control to compare samples of green leaf material and soils that were collected by Nadelhoffer and coworkers in 1999 and analyzed by the same analytical techniques to determine the effects of heavy N fertilization. Additionally, samples collected in 1999 from control, low N and high N plots were compared to separate changes associated with forest changes and N fertilization.

13C NMR spectra from all samples contain predominantly signals associated with paraffinic (0-45ppm) and carbohydrate (60-90 ppm) structures and low amounts of aromatic (90-160 ppm) structures. ¹⁵N NMR spectra reveal signals associated with amide structures. The mineral soil samples exhibit minor amounts of aromatic structures, suggesting little contribution of lignin structures to the organic matter. Lignin added to the soil by plant inputs may have been oxidized, degraded and/or leached out. Primary changes observed in each set of the canopy-to-soil profiles were a loss of carbohydrates and a persistence of the paraffinic carbon regions of the NMR spectra (Fig. 1 and 3).

Of particular interest are the effects of N fertilization on leaf growth and litter decomposition in the various soil horizons. Increased levels of methylene structures (30 and 32 ppm) are observed in the NMR spectra of black oak green leaf samples with increased fertilization (Fig. 1). However, most of these additional structures are removed by lipid extraction (Fig. 2) indicating that the structures may be associated with easily removable/degradable components such as chlorophyll and other extractable lipid compounds. Increased levels of lignin (90-160 ppm) and aliphatic amide carbons (160-190 ppm) are also observed in the NMR spectra of the leaf samples from high N plots relative to control samples (Fig. 1). Spectra of soil samples collected from both hardwood and pine high N plots from the Oe and Oa horizons reveal greater amounts of carbohydrate, lignin and aliphatic amide carbons relative to paraffinic carbons (Hardwood shown, Fig. 3). This indicates that the rate of decomposition has decreased with increased N fertilization, supporting results of a litter decomposition study performed by Magill and Aber (1998). Pyrolysis GC/MS results (in progress) provide additional details such as the presence of phytadienes, derived from chlorophyll and aid in the understanding of changes that are occurring in the Chronic N plots.

Magill, A.H. and Aber, J.D. 1998. Plant and Soil 203: 301-311.

---

Relationships Among Woodland Vegetation and Land-Use History, Cape Cod National Seashore, MA
R. Eberhardt, D. Foster, G. Motzkin, J. Harrod and B. Hall

Human land-use affects the composition, structure, and landscape pattern of vegetation worldwide, and human disturbance history must be considered when interpreting modern vegetation and developing conservation objectives and strategies. The specific effects of land-use history on modern vegetation are poorly understood because land-use history is complex, poorly documented and often correlated with environmental gradients. In this study we are examining relationships among modern woodland vegetation and land-use history on 5100 ha of glacial outwash deposits on Cape Cod National Seashore in southeastern Massachusetts. Like previously-studied sand plains in the Connecticut Valley (Motzkin, et al. 1996, 1999), relatively homogeneous site conditions and the availability of detailed historical information make the study area particularly appropriate for evaluating the effects of land-use history on vegetation. The conservation significance of the site also provides an opportunity to inform management efforts.

Colonists settled outer Cape Cod by 1644, and US Coast and Geodetic Survey maps suggest that agriculture/open land covered approximately 48 % of the study area during the mid-nineteenth century. Land abandonment and reforestation followed: based on 1938 aerial photographs, agriculture/open land covered approximately 15 % of the study area by the mid-twentieth century. Using historical maps and the presence/absence of plow (Ap) horizons, we have grouped randomly-selected modern vegetation plots (n = 91) into four broad categories of historical land-use (cultivated/improved pasture, unimproved pasture/open, woodlot, and military use). Overall variation in species composition relates significantly to variation in land-use history (Fig. 1), with common understory species occurring preferentially in certain land-use categories (Fig. 2). Preliminary results indicate that Deschampsia flexuosa, Arctostaphylos uva-ursi, Prunus serotina, Amelanchier spp. and Trientalis borealis occur more frequently in former croplands or improved pastures, Smilax rotundifolia and Myrica pensylvanica concentrate in former umimproved pastures or open areas, and clonal ericaceous shrubs favor former woodlots. This study illustrates that land-use history can exert a dominant effect on modern vegetation in coastal New England and must be considered when developing conservation objectives and management guidelines.

Motzkin, G., D. Foster, A. Allen, J. Harrod, and R. Boone. 1996. Controlling site to evaluate history: vegetation patterns of a New England sand plain. Ecol. Mon. 66: 345-365

Motzkin, G., W.A. Patterson III, D.R. Foster. 1999. A historical perspective on pitch pine-scrub oak communities in the Connecticut Valley of Massachusetts. Ecosystems 2: 255-273

---

Songbird Nest Predation in Eastern Oak Forest: Variations with Nest Height
R. Field

Predation is a major cause of nest failure for many open-nesting forest bird species. Understanding the ecology of nest predation is important for conservation and management of forest birds. Although songbirds nest throughout the vertical layers of a forest, little is known about predation in the higher strata. Research on nest predation at different heights of vegetation has been limited to ground or shrub levels. Few studies have examined predation events at bird nests above 2-3 m.

One potential predator, the white-footed mouse (Peromyscus leucopus), is omnipresent in the mixed oak forests and a capable climber. However, we know little about mice as predators of songbird nests, especially those in the canopy and subcanopy. Pervious studies of nest predation often used artificial nests baited with quail eggs. Because mice have difficulty depredating the relatively large quail eggs, their role as a potential predator throughout the forest may have been underestimated.

In the summer of 1999, I continued a study of predation of songbird nests in the mixed oak stands of Harvard Forest and the northern Quabbin Reservation. The objectives of this study are to estimate nest predation rates in the vertical layers of a mixed oak-forest and evaluate the role of mice as potential predators of those nests. The work is being done in parallel with C. A. Langtimm’s study of population estimates and vertical distribution of P. leucopus.

In 1997, I conducted point count at 2 sites on Prospect Hill area of Harvard Forest. I identified 35 species of birds on the study plots, 25 of which were known or probable breeders. Among those were species that tend to nest on the ground, in shrubs or low branches, in subcanopy trees, and higher into the canopy. In 1998 and 1999, I expanded the study to sites at the Tom Swamp forest and the north Quabbin Reservation (near Gate 19). With the help of summer technicians from the Harvard Forest program, Research Experience for Undergraduates in both 1998 and 1999, we conducted bird surveys at all three sites (1998 only), and artificial nest experiments using plasticine eggs. Eggs made of modeling clay or plasticine have been used in nest predation studies to document potential predators based on tooth or claw marks on the eggs. Nests were placed at sites on the ground, in shrubs, in the low subcanopy (mean height of 7.4 m), and in higher subcanopy (mean height 11.3 m). In 1999, I repeated the artificial nest experiments using ground-shrub grids at the same 3 sites, but doubled the number of tree nest sites.
Results are preliminary at this time because the project will continue in the summer of 2000. The importance of a long-term study was apparent in the 1998 and 1999 data because of the variability between years. Overall nest depredation was lower in 1999 compared to 1998, with less evidence of mouse depredation at the artificial nests (Table 1). In 1998, nest predation was lowest in Tom Swamp compared to Prospect Hill and the Quabbin site. However, in 1999, depredation tended to be higher in Tom Swamp than the other 2 sites. Although numbers of depredated nests were lower in 1999, there was a similar pattern in predation rates for the 4 strata (Table 1). Ground nest predation was lowest, shrub and low subcanopy nest predation were similar, and high subcanopy nest predation was highest.

Width of incisor marks in the "depredated" plasticine eggs for both years were variable in the ground and shrub nests, indicating a diversity species. Although sample sizes were small for the higher nests, the eggs in the subcanopy were marked with only smaller incisor indentations in the range of tooth marks from museum samples for Peromyscus.

The lower predation rates in 1999 may have been the result of a lower population of mice in 1999 and differences in precipitation between the 2 years. Because of abundant acorns in the fall of 1999, we anticipate an increase in the mouse population in 2000 that will make the nest predation results more comparable with the 1998 season. For the 2000 season, I will repeat the artificial nest experiments for comparison with the previous 2 years.

The information from this study will help us better understand predatory risks for songbirds nesting throughout the vertical structure of New England oak forests.

---

Forest-Atmosphere Exchange Processes: Report on Activities 1999-2000
D. Fitzjarrald, K. Moore, R. Sakai, J. Freedman, O. Acevedo, R. Staebler, G. Wocjik, and M. Czikowsky

1. Introduction

We have been studying how mass and energy exchanges between the forest and the atmosphere occur, and what consequences they have on the local microclimate. Our efforts have been motivated by the following questions about flux measurements:

a) If the eddy covariance method is so good, why is the energy balance not closed?

b) Are flux and mean profiles similar across different forest canopy types?

c) Does the presence of topographically controlled local winds bias flux observations?

d) What really goes on at night (as regards CO₂ flux observations)?

Our second effort is motivated by curiosity about how representative point flux measurements can be to the wider region, and to what degree do forests alter climate:

e) What changes occur in the lower atmosphere when leaves emerge?

f) Does the occurrence of forced cumulus clouds play a pivotal role in reducing stress to the vegetation?

2. Energy balance /turbulence studies

Similarity forms in the canopy and roughness sublayers

Based on data from the EMS tower and other sites we have quantified the shape of heat, water vapor, and CO₂ flux cospectra in the roughness sublayer, just above the canopy. These are not identical to the commonly cited forms found over flat terrain. Consequently, fluxes calculated using fixed averaging intervals (the common approach) "miss" an important low-frequency contribution, but the missing flux can be estimated using a correction that depends on mean wind speed.

Previously we showed that the drag coefficient changes little as the Harvard Forest goes from leafless to the foliated state. A new result is that the profiles of s_w and u_* (friction velocity) are similar within the canopy for each state, if a foliage-weighted canopy thickness z_c, which depends on the profile of vegetation in the canopy, is used for the vertical scale. This similarity extends to a variety of broadleaf forests; a different but still similar profile occurs for conifers (Fig. 1). There is a general expression for the displacement d/h = 0.38 z_c for broadleaf forests, and this replaces the commonly used d/h = 0.75, where h is canopy top height.

DRAINO-99. On calm nights, many observers find that local CO₂ buildup in the canopy, even when combined with the direct flux measurement, lead to unrealistically low estimates of ecosystem respiration. This problem may be the result of averaging the eddy flux improperly, or as Lee (1998) is the most recent to suggest, the local drainage flows may advect the "missing" CO₂ away. Drainage circulations are notoriously difficult to observe, and they rarely occur as sustained flows. Such flows at the forest floor are further complicated by the presence of tree trunks and subcanopy vegetation. In DRAINO-99, we deployed instruments at the EMS for four months to observe subcanopy flows. Five sonic anemometers (four 2-D and one 3-D) were deployed at 1.5 m around the EMS tower. Fast-response observations of [CO₂] and specific humidity were made with the 3-D anemometer. An acoustic sounder operating at canopy height from a nearby scaffolding tower reported wind speeds and turbulence intensity at 25m intervals up to 750m. Ongoing continuous measurements at the EMS tower include those from 3-D sonic anemometers at 30m, 18m and 11m, the radiative budget components, and temperature, CO₂ and H₂O profiles. On two nights, we observed the motion of neutrally buoyant bubbles along the forest floor. Preliminary results include: a) Understory flows are most frequently decoupled from the flow above (see "snapshot", Fig. 2); b) On average, nighttime flows are from Prospect hill or the nameless hill to the west of it. From the bubble observations, we observed that these topographic flows are not sustained; c) The diurnal cycle of understory wind speeds peaks in the early afternoon, in phase with temperature. We are currently studying the occurrence of gravity waves inside the canopy and in the stable boundary layer above at night. In future, analysis of the acoustic sounder record will also allow us to model the effect of Prospect Hill on HF fluxes.

3. Climatalogical Studies

Climatological studies. At Worcester MA, we calculated R, a relative humidity and B', a tendency Bowen ratio, based on T_max and q, the daily average specific humidity. R goes through a minimum near the time of leaf emergence at Day of year 116. At extremes of R, ðR/ðt = 0, and Lðq/ðt = R e(C_p ðT_maxx/ðt), and it follows that B' = 1/(R e), where e = [L/C_p]ðq_sat/ðT, with q_sat(T, p) the saturation specific humidity at given temperature and pressure. B = 1/e is the "equilibrium" flux Bowen ratio. Previously it has been argued that evaporation only proceeds at near to the equilibrium rate because fluxes are converging beneath a rigid lid, but we found that only 5% of surface fluxes are trapped in the mixed layer in long-term averages. There is another mechanism to maintain R nearly steady during the growing season. The agent seems to be the appearance of boundary layer cumulus clouds (BlCu), a result that came from analysis of a composite of case studies.

Composite study. Following a typical frontal passage, the mixed layer (sfc to 1500 m) functions as a heat and moisture reservoir, as local and regional warming and moistening modify the overlying air mass. The presence of forests locally modifies these air masses through evapotranspiration (ET). Increased BL humidity reduces the vapor pressure deficit (VPD) and surface temperature, and this facilitates the formation of Blcu. A composite of the BLcu sequences showed a rapid drying and cooling of the mixed layer followed by a gradual warming and moistening over a period of 4-5 days. Budgets of the temperature and humidity tendencies showed that, during days 2-5 of the sequence, the mixed layer (ML) grows rapidly into the previous days' residual layer. The entrainment terms (heat and moisture) contribute only 10-20% to the daily ML heat and moisture tendency after the first day. The presence of active cumulus clouds modulates the exchange of mass, heat, and moisture between the ML and free atmosphere above, maintaining an approximate mixed layer equilibrium at constant R.

Lee, X. 1998. On micrometeorological observations of surface-air exchange over tall vegetation. Ag. Forest. Met. 91: 39-49.

---

A New Approach to Modeling Soil Carbon Dynamics
S. Frolking

Many soil biogeochemistry models (e.g. Century, DNDC, etc.) are obliged to partition total soil organic matter (SOM) into a series of pools (e.g., fast, slow, passive), without a sound empirical methodology for doing so. This is further complicated if the models attempt to develop a vertical profile of SOM.

I am working on a new approach to this problem, building on SOM modeling work I have done for boreal forests and peatlands. In this approach, annual litter fall is deposited into a litter cohort. This cohort then loses mass, with its decomposability declining linearly with mass lost

(1)

where m₀ is the litterfall, and k₀ is the initial mass loss rate for this tissue. This approach was generally successful in modeling SOM accumulation over millennia in peatlands, where annual cohorts are well-stratified. However, for forested ecosystems, litter does not remain stratified, and, when root litter inputs are considered, doesn’t even begin that way. To account for this, the litter cohorts (combined above- and below-ground) will have a prescribed initial distribution, and will disperse/diffuse through the soil profile with time. Since modeling actual diffusion of SOM would involve so many unknowns (for which there are so few data) it will instead be prescribed in what may be a reasonable form. The function I am currently exploring for the SOM profile is the gamma distribution (Fig. 1a)

(2)

where z is depth, and a and b are parameters. b corresponds to the decay in SOC content down the soil profile, and a relates to the skewness of the distribution. Combining these two ideas (decomposition and dispersion), the SOM would be described by a collection of annual cohorts with the following mass distribution (Fig. 1b)

(3)

The total mass of any annual cohort would be given by integrating this over the depth profile and would be as in equation (1). The total SOM at any depth would be given by adding all the cohort contributions at that depth. Time evolution of the SOM distribution will be generated by making a and b functions of time.

This approach is mathematically tractable, and can be explored in very simple models initially. The model will generate profiles of SOM, decomposition, CO₂ production, and can easily be ‘labeled’ to generate D¹⁴C values for SOM and CO₂ production, all of which can be tested against observation.

---

Carbon Exchange by an Old-Growth Hemlock Forest in the Summer of 1999: Eddy Covariance Data Compared to Model Predictions
J. Hadley and E. Huber

We measured carbon exchange near the NE corner of a ~200 year old hemlock stand using eddy covariance during twelve days in mid-summer 1999. Wind was primarily from the SW, or from the main part of the hemlock stand, during 6 to 12 hours in each of three days and four nights. During these periods the average measured nocturnal ecosystem respiration (R_e) was about 4.8 µmol m^-2 s^-1. Adjustments for CO₂ storage during calm periods raised this to about 6.2 µmol m^-2 s^-1. An ecosystem carbon exchange model estimated R_e at about 7.3 µmol m^-2 s^-1 during these periods, or 17% above the measured value. An overestimate of surface soil water content could explain some of this difference. Soil water in the top 5 cm of soil was unusually low at the beginning and end of eddy covariance measurements, but two small rain events occurred in the interval, and the model estimate of soil respiration was very sensitive to surface soil moisture. Soil respiration was about 65% of estimated R_e. Overestimates of leaf area or sapwood volume could also have contributed to the apparent 17% model overestimate of R_e.

Eddy covariance showed that R_e was higher in the hemlock forest than in surrounding hardwood forests. During the twelve-day measurement period, R_e was above 5 µmol m^-2 s^-1 during 46 % (19 of 41) half-hourly nocturnal periods with wind from the SW quadrant and adequate turbulence for eddy covariance measurements. In contrast, R_e was above 5 µmol m^-2 s^-1 only 7% of the time (6 of 82 measurements) with other wind directions. Average R_e with SW wind was more than twice Re with wind from the SE and NW quadrants (Fig. 1, p< 0.01). Average R_e during periods of NE wind was intermediate (3.8 µmol m^-2 s^-1), but this included two of three extreme outliers in the data set. Without these points, R_e from the SW quadrant was over twice R_e from all other directions.(p<= 0.002). Wind speed, momentum flux, air temperature and time of night all had small effects on R_e, but with these accounted for statistically, average R_e from the SW was still higher than from any other direction, and was greater than for SE and NW at p< 0.05.

During daylight, eddy covariance measured average C uptake of the hemlock stand at 6.3 µmol m^-2 s^-1, but the carbon exchange model predicted only 5.0 µmol m^-2 s^-1. The difference between measured and modeled C flux in the daytime (1.3 µmol m^-2 s^-1) was very close to the nighttime difference, so an overestimate of R_e could explain both differences. Weighting the average day and night fluxes by the length of light and dark periods (~14 and 10 hours) gave an average measured C uptake of 1.1 µmol m^-2 s^-1 (1.14 g m^-2 d^-1) and a very small model-estimated C release of 0.09 µmol m^-2 s^-1 (0.09 g m^-2 d^-1).

---

The Forest Vegetation of Cape Cod: a Regional-Historical Analysis
J. Harrod, D. Foster, G. Motzkin, B. Hall and R. Eberhardt

Our study documents changes in land use and forest cover in Cape Cod, MA over the past 400 years and evaluates influences of historical land use, regional position, and geologic substrate on modern forest composition and structure at landscape to regional scales. Pollen and witness tree data suggest that the Cape was largely forested and that pitch pine and oaks were the most abundant trees at the time of European settlement in the 17th century. Fuelwood cutting and land clearance led to a rapid loss of forest cover; by the 18th C, wood shortages and severe soil erosion occurred in several towns. A landcover map from the mid-19th C shows the Cape as 49% forested, with most areas near the coast cleared and large patches of forest remaining in the interior (Fig. 1). Forest cover increased following farm abandonment and afforestation efforts, exceeding 65% in 1951, but has since decreased due to residential and commercial development, falling below 50% by 1990.

Our analysis of the modern forest vegetation, based on 270 plots, indicates that Cape-wide vegetation patterns reflect historical land use, geologic substrate, and regional position. Several species show strong spatial variation, with mesophytic trees and vines most abundant on the inner Cape and early-successional shrubs and grasses most abundant on the outer Cape (Fig. 2). Within landscapes comprised of single geologic substrates, history of past land use is the principal factor driving compositional patterns (Fig. 3). While 400 years of land use have dramatically transformed the Cape's land cover and strongly influenced distributions of individual species, modern canopy composition is strikingly similar to that inferred from 17th century witness tree data.

---

Differential Responses of Trees to Drought During Ontogeny in New England: a Species-By-Ontogeny Interaction
J. -S. He and F. Bazzaz

We developed a framework for scaling seedling responses to mature trees combining both environmental and ontogenetic differences between seedling and adults. We investigated how radial growth of juveniles and mature trees differ in responses to drought in temperate tree species in New England. Saplings and mature trees of four species, which ranked according to increasing shade-tolerance (successional position) as paper birch < red maple <yellow birch <sugar maple, were sampled in well drained Canton loam soil at Harvard Forest.

Mature trees experienced differential ring-width reduction during four severe drought periods occurring from 1957 to 1995. Except for 1957, the four species have not any significant difference in their relative changes in radial growth during drought, but their relationships between ring-width index (RWI) and monthly Palmer Drought Severity Index (PDI) were varied. The late successional species (sugar maple) and mid successional species (yellow birch) had significant positive correlation with several monthly PDI in current and previous years. In contrast, early successional species (paper birch) and mid successional species (red maple) showed no significant positive correlation with monthly PDI in current or previous year. All saplings of all species exhibited significant growth reduction during 1995 drought, and showed significant interspecies difference. A relative change in radial growth of a late successional species (sugar maple) was more significant than that of the other 3 species in saplings during drought 1995 (Fig. 1). No saplings of the species showed positive correlation between RWI and monthly mean PDI of previous year. A three-way (species ´ ontogeny ´ sites) ANOVA analysis showed that the differences between saplings and mature trees in paper birch and sugar maple were significant, showing a species-by-ontogeny interaction. The result that saplings of paper birch had a less, and saplings of sugar maple had a bigger growth reduction than mature individuals suggested there is an ontogenetic shift as saplings grow to mature individuals.

---

Atmospheric Reactive Nitrogen at Harvard Forest: Measurements in Progress
C. Horii, M. Zahniser, J. Munger and S. Wofsy

Semi-continuous measurements of nitric acid (HNO₃) and nitrogen dioxide (NO₂) began at the Harvard Forest Long-Term Ecological Research site in August 1999. A dual-channel tunable diode laser absorption spectrometer (TDLAS) provides the high sensitivity (100 ppt) and fast time response (1 second) needed to record ambient variations and deposition rates by the eddy covariance method (for instrument details, see Horii et al. 1999). The new measurements augment the existing suite of trace gas concentration and flux measurements at the site and facilitate the detailed study of reactive nitrogen (NO_y) speciation and effects on ozone (O₃) production over seasonal and annual time scales. The measurements will continue for at least a year, providing the necessary data to assess the relative importance of HNO₃ deposition as a sink for tropospheric reactive nitrogen compared to NO₂ and Peroxyacetylnitrate (PAN). A separate abstract by Munger et al. (2000) presents PAN measurements. Monitored species at the site include NO_y and O₃ concentrations and dry deposition fluxes, NO_x (NO + NO₂) and PAN concentrations, non-methane hydrocarbons, and other tracers of anthropogenic pollution such as CO, CO₂, CFCs, and SF₆.

Deployment of the TDLAS instrument for unattended, continuous, outdoor operation in all seasons represents a significant breakthrough in the use of what was once a laboratory-based technique. TDLAS instruments have been used successfully at field facilities where air samples may be drawn into an indoor laboratory, outdoors for short measurement campaigns in fair weather, in mobile laboratories for limited periods, and on aircraft. The dual TDLAS at Harvard Forest has been installed in two weatherproof, temperature-regulated enclosures on the top two stages of a scaffolding tower, roughly 100 m from the main EMS tower. Maintenance includes bi-weekly retrieval of data and transport of 60 liters of liquid nitrogen to the site in order to cool laser diodes and detectors, periodic replacement of compressed nitrogen gas cylinders, and occasional refilling of trace gas reference cells. Throughout the fall season, the primary elements of the system have continued to function at the expected sensitivity levels. Work continues during the winter and spring of 2000 to reduce data dropout periods, and increase concentration and flux data coverage for both HNO₃ and NO₂.

A detailed comparison of TDLAS and slow Photolysis/O₃-Chemiluminescence NO₂ measurements is currently being conducted. This will include measurement of the Photolysis/O₃-Chemiluminescence reference gases by the TDLAS to crosscheck calibration values. Analysis of continuous 1-second NO₂ data from the TDLAS instrument will then be analyzed with 8 Hz vertical winds from the sonic anemometer to produce dry deposition flux values. The technique will then be extended to HNO₃ measurements.

Figures 1 and 2 show examples of HNO₃ and NO₂ spectra from the dual TDLAS obtained during the fall at Harvard Forest. The unambiguous detection of the particular species is evident in the match between the measured spectra and the non-linear least squares fits to tabulated spectral parameters. Longer time-averaged spectra are presented to show high signal-to-noise examples, but the usual mode of operation uses a fast spectrum accumulation time of 1 second.

Horii, C. V., M. S. Zahniser, D. D. Nelson, J. B. McManus, S. C. Wofsy. 1999. Nitric Acid and Nitrogen Dioxide Flux Measurements: a New Application of Tunable Diode Laser Absorption Spectroscopy, SPIE 3758: 152-1651.

Munger, J. W., C. V. Horii, S. C. Wofsy. 2000. Biogenic hydrocarbons and nitrogen oxides: A forest atmosphere interaction with implication for regional air pollution, 11^th Annual Harvard Forest Ecology Symposium Abstracts.

---

Timber Harvest as a Form of Disturbance in the North Quabbin Region
D. Kittredge and D. Foster

In addition to numerous sources of natural disturbance acting on the forested landscape of southern New England, the commercial harvest of timber creates disturbance effects by removing parts of trees, reducing stand density and elevating light levels, disturbing mineral soil, and altering stand structures. Knowledge of the extent to which this form of disturbance occurs, and the typical size, frequency, intensity, and pattern of such disturbance, will contribute to an improved understanding of ecosystem dynamics in this forested region.

The commercial harvest of timber (i.e., 25 Mbf (59 m³) or more from an operation) is regulated in Massachusetts through the Forest Cutting Practices Act. An approved permit is required to conduct such an operation. These permits require information on the volume of wood removed, as well as environmental protection features to be applied during the operation. The permit also requires a map depicting the spatial extent of harvest. These permits provided a data source for our study, which seeks to depict the extent and pattern of harvest as a form of disturbance in the 19-town North Quabbin region (168,300 ha). Harvest permits were reviewed and tabular and spatial data were captured on each operation since 1984.

Over the course of 16 years, preliminary data analysis indicates: 1,992 commercial harvests; average area = 16.8 ha (maximum = 338.7 ha, minimum = 0.15 ha, sd = 19.1 ha). During this time period, commercial harvest covered 33,404 ha (an average of 2,091 ha/year, in 137 operations), representing 23.3% of all forest in the north Quabbin region (Fig 1). During this time, 3,349 ha had been harvested more than once. Mean harvest intensity was 44.8 m³/ha (7.7 Mbf/acre).

We intend to continue capturing these data over time, and to further analyze data looking for landscape patterns of this disturbance, as well as where this disturbance is not happening. These data contribute to an improved understanding of ecosystem dynamics, and further could be used in studies of carbon flux in this landscape.

---

Assessing the Indirect Consequences of the Hemlock Woolly Adelgid: Ecosystem and Vegetation Response to Hemlock Logging
M. Kizlinski, D. Foster and D. Orwig

Hemlock woolly adelgid (HWA) (Adelges tsugae), an exotic aphid-like insect from Japan, was introduced to southern Connecticut in the mid-1980s and has since spread to central Massachusetts and continues to move northward. With no effective native predators or host resistance, HWA has left thousands of hectares of dead or dying eastern hemlock (Tsuga canadensis) forest. Landowners are now required to make management decisions regarding the fate of their hemlock, decisions that have not been necessary in the past. The option of logging, either before the arrival of HWA or during the early stages of infestation, presents a series of questions that have not been formally addressed. Eastern hemlock has not been a commercially valuable species, so information on the silviculture and management of the species and the impact of logging on ecosystem and vegetation dynamics is neither extensive nor complete. This project will address these issues, which are growing in importance as more and more hemlock stands are logged in response to HWA.

Fifteen logged stands were chosen for this study to cover the range of ages and successional states since HWA arrival and hemlock logging commenced about 15 years ago. Site selection was limited to those locations that were dominated by hemlock, were heavily logged, and are located on comparable soils. These sites represent 13 years of post-logging succession and ecosystem recovery.

Vegetation plots have been established in eight stands and data has been collected on pre-harvest stand composition and basal area, intensity of harvest, and sapling (>2 cm diameter at breast height (dbh)) density. This summer, these plots will be revisited and data on seedling and herbaceous cover will be collected. Ecosystem parameters to be measured this summer include nitrogen (NH₄⁺ and NO₃^-) availability and mineralization rates, decomposition, pH, bulk density, moisture, and soil organic matter content.

By comparing the results of this study to other ongoing HWA work at the Harvard Forest, comparisons can be made between the effects of logging and ‘natural’ mortality of hemlock stands. Using this information, better informed management decisions can be made regarding newly infested stands, as well as those areas that face imminent infestation.

---

Variation in Vertical Activity of White-footed Mice in Oak Forest
C. Langtimm

A third year of study of the climbing patterns of white-footed mice, Peromyscus leucopus, revealed substantial annual variation in the vertical activity of this species in understory trees and shrubs. The proportions of arboreal captures at a moist site on Prospect Hill was 59% and 64% in July and September of 1997, but in June and August of 1998 proportions dropped below 50% to 37% and 44%, and then dropped even further in June and August 1999 to only 14% and 23%. Estimates of population abundance were comparable in 1997 and 1999 and do not explain the differences in proportions between those years. 1999, however, was an extremely dry year compared to the previous years and may have had an impact on prey items accessible to mice in shrubs and trees. Arboreal captures at a more xeric site in Tom Swamp were similar in 1998 and 1999 (29% and 30% in June and August 1998 and 25% and 33% in June and August 1999), supporting the likelihood of site-specific differences in climbing activity for the species. Limited trapping in the crowns of canopy oak trees continued to document the presence of white-footed mice at higher levels of the forest, although fewer individuals were captured during the summer of 1999 than in the previous two years. The documented seasonal, annual, and site-specific variation in climbing activity of this species demonstrates the behavioral plasticity of white-footed mice to utilize all levels of the forest. This plasticity could have important implications for our understanding of their role in the epidemiology of Lyme disease, gypsy moth outbreaks, and songbird nest predation.

---

Does Carbon Quality Limit Soil Carbon Loss Due to Warming?
H. Lux, T. Ahrens, J. Melillo, P. Steudler, A. Ricca and F. Bowles

The soil warming experiment, designed to investigate the effects of a 5ºC increase in temperature on soil processes, has been running since the summer of 1991. Over the past 9 years, CO₂ fluxes, which represent both root and microbial respiration, have been measured from April to November in heated, control, and disturbance control (DC) plots. Extensive sampling of the soil organic and mineral horizons was undertaken in 1999 to investigate soil carbon stocks. Through the examination of these data, along with the application of our knowledge of carbon cycling dynamics across ecosystems, we are refining our hypotheses of soil responses to warming.

Long-term pattern of soil respiration in response to warming – During the first five years warming increased the average annual rate of soil respiration by about 20%. After 1995, the CO₂ efflux differences between the heated plots and the disturbance controls began to decline until there were no measurable differences in 1999 (Fig. 1).

Relative effects of warming on root and microbial respiration - In a separate warming experiment, designed to investigate the relative contributions of root and microbial respiration to total soil respiration, we have consistently observed that about 20% of CO₂ efflux is contributed by root respiration, and about 80% of CO₂ efflux is the product of microbial respiration.

Estimating carbon losses from the soil in response to warming – Combining information from the CO₂ efflux measurements and the study to separate root from microbial respiration, we have estimated that 934 g C/m² have been lost from the soil over the first 9 years of the experiment (Fig. 2).

Comparison of soil carbon stocks in heated and DC plots in 1999 allows us to quantify carbon loss from the soils directly. Thus far we have completed this analysis for the organic horizon, and from these direct measurements we calculate a loss of carbon from the O horizon of 260g C/m². The carbon lost from this horizon is about 20% of its total stock.

A two-pool soil carbon concept and the potential for positive feedback – Based on results from the soil warming study, we hypothesize that there are at least two pools of soil carbon at the Harvard Forest: an easily decomposable pool rich in relatively simple carbon compounds that are readily accessible to microorganisms; and a difficult to decompose pool made up of complex ring compounds and otherwise lignified carbon. The decline in the difference between CO₂ released from the heated and DC plots in the years following 1995 may represent the exhaustion of the readily accessible pool of carbon in the heated plots. Our study causes us to suggest that while CO₂ efflux from mid-latitude forests will lose carbon in response to warming, a long-term and large positive feedback response to warming is not likely.

---

Chronic Nitrogen Additions to Two Forest Stands
A. Magill and J. Aber

The nitrogen saturation potential of temperate forests in the Northeastern United States and in Europe is of mounting concern as nitrogen deposition from anthropogenic sources continues to increase. In order to determine the long-term impact of N inputs, key indicators of nitrogen saturation, such as forest productivity, foliar chemistry, soil net mineralization rates and soil solution chemistry, have been measured on plots receiving NH₄NO₃ additions at the rate of 0, 5 and 15 g N m^-2 yr^-1 since 1988. A nitrogen plus sulfur treatment was also included through 1998 with NH₄NO₃ additions at the same rate as the low N plots plus 7.4 g N m^-2 yr^-1 of sodium sulfate. Beginning in 1999, the N+S plots received only the NH₄NO₃ additions. Plots were established in two forest stands, a red pine plantation and a mixed hardwood stand, and are located in the Prospect Hill tract of Harvard Forest.

Increased soil water nitrate concentrations have been measured in the pine high N plot since 1989 and in the hardwood high N plot since 1995. In 1999, pine high N mean annual nitrate concentration was 20-fold greater than pine control; hardwood high N mean annual nitrate concentration was nearly 10-fold greater than control. Green foliage N content of the dominant species has increased dramatically with nitrogen treatments. Percent N in foliage from the high treatment plots has been measured at 80% and 25% greater than control foliage for red pine (pine stand) and black oak (hardwood stand), respectively. Changes in foliar biomass have not been significantly different overall.

Data currently under analysis include aboveground woody biomass, root biomass, root nitrogen and carbon content, soil bulk density, fungal and microbial biomass. Canopy access towers were installed in 1999 and in situ photosynthesis measurements are planned for the 2000 field season.

---

Embolism Intolerance of Black Willow (Salix nigra Marshall)
P. Melcher, M. Zwieniecki and N. Holbrook

Parameters related to water transport properties of 3 year old potted black willow (Salix nigra L.) plants growing in a glasshouse located at Harvard Forest were studied from June to August of 1999 to address current issues surrounding embolism repair. It has been previously demonstrated in other species that the ratio of liquid to gas filled vessels changes diurnally, suggesting that refilling of cavitated vessels can occur during times when neighboring vessels are under tension. However, it is not clear if these "repaired" vessels reestablish complete hydraulic continuity with neighboring vessels (e.g., liquid-liquid contact across pit pores) making them truly "functional" in long distance water transport. We found that experimentally induced xylem sap tensions of 1.5 MPa reduced the hydraulic conductivity (k_h) by 50% for black willows. To investigate the potential of vessel refilling during times of tension driven sap flow we made concurrent measurements of stem specific k_h and leaf water potential (Y_L). No diurnal changes in stem specific k_h were observed. This is most likely a result of the fact that Y_L never reached values lower than 1.5 MPa throughout the day. The effect of air seeding (the minimum pressure required to force gas across the small diameter pit pores) on whole plant transpiration (measured using a balance) was determined by applying gas pressures across the main stems (base of plant) of intact potted black willows using pressure collars attached to a high pressure gas delivery system. We found that pressures less than 1.5 MPa across the pit-pore gas-water interface (applied pressure plus leaf balance pressure) had no effect on whole plant transpiration rates (Fig. 1); however, when the pressure across the air-water interface of the pit pores were greater than 1.5 MPa, transpiration was dramatically reduced followed shortly by leaf wilting and eventual leaf senescence. This indicates that the critical air seeding pressure is similar to tensions that reduce k_h by 50% and the non-restoration of the transpiration gives supporting evidence that embolism formation is catastrophic for black willow plants. This contrasts with recent observations that many species tolerate and repair cavitation on a diurnal basis. These findings give implications that different strategies for maintenance of the continuity of the hydraulic pipeline from soil to leaf exist and provide insight on issues surrounding evolutionary adaptations, plant distribution, and species sensitivity to drought.

Acknowledgements

We are grateful to Emily Huhn and Barbara Muñoz for assistance in the laboratory and the field.

---

Retrenching at the Harvard Forest DIRT Plots
P. Micks and K. Nadelhoffer

Our main focus on the DIRT experiment in 1999 was the accomplishment of major mainentance work on the two root-exclusion treaments first installed in 1990, along with the other experimental manipulations that comprise the DIRT project. In that year, we began a long-term study of the linkages between forest soil organic matter dynamics and inputs of above- and below-ground litter (DIRT: Detritus Input Removal and Transfer). The goal of the DIRT project 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. Treatments include: doubling aboveground litter (double litter), exclusion of aboveground litter (no litter), exclusion of root inputs by trenching (no roots), and exclusion of aboveground litter and root inputs (no inputs), on replicated 3m x 3m. plots (n=3). An additional treatment (O/A-less), implemented in 1991, replaced O and A horizon material with B horizon soil, with normal litter inputs allowed to occur thereafter.

The most important finding from the first eight years of the project was the dominant influence of root inputs over above-ground litter inputs on soil respiration. Similar CO₂ efflux patterns were demonstrated in laboratory-incubated soils taken from the plots in the fifth year of treatments. Relative differences in CO₂ flux were proportional to aboveground litter inputs when roots were intact, although by the eighth year these differences were less clear.

Because of concern that tree roots had reinvaded the root-excluded plots, in August of 1999 we undertook the major task of redigging the trenches around all of these plots and installing new root barriers. Trenches were dug to the original 1 m depth and the original rigid plastic barrier material was removed. We noted that in a few cases roots had penetrated the barrier, but most roots invaded by growing underneath the barrier, despite the fact that root density is very low at those depths in the DIRT study area. Trenches were relined with a combination of heavy-duty nylon-reinforced plastic and polyethylene sheeting, then backfilled with the excavated soil. We expect that retrenching will be necessary every 5-8 years. We anticipate continuing the DIRT experiment for many more years, and this work ensures that we will continue to gain valuable data about the long-term role of plant litter in forest soil organic matter dynamics.

---

A Regional Climatology of Cloud and Aerosol for Forest-Atmosphere Exchange
Q. Min and K. Moore

Clouds and aerosols affect the forest light environment by enhancing diffuse PAR, or, in the case of optically thick clouds, reflecting most irradiance into the atmosphere. Our efforts are to integrate multiple existing data sets, including surface and satellite measurements (listed in Table 1), to develop a regional climatology of cloud, aerosol and ozone to be applied to forest-atmosphere energy exchange, and to improve satellite-based measurements. Combining various data sets for processing, we currently focus on two sites: Albany, NY the station with the longest record, and Harvard Forest where surface exchange and radiation are measured. Further, we have used one year's worth of satellite data (GOES 10, 1998) to study spatial and temporal effects of cloud and aerosol. Sample recent results are described below.

Optical depth of aerosol and cloud and its spectral dependence are important factors in the total short wave, PAR, and UVB irradiances. Figure 1 is a six-year time series of aerosol optical depth at 415 nm and Angstrom exponent coefficient from MFRSR at the Albany site. The Angstrom coefficient is related to the aerosol particle size distribution: the smaller Angstrom coefficient, the bigger aerosol particle size. These data show the clearing of the stratosphere following the 1991 eruption of Mt. Pinatubo, the excursion of the Pinatubo event above the apparent local annual cycle and recovery of the stratosphere to "background" levels in magnitude as well as in particle size distribution. The annual cycle of optical depth for aerosol and thin cloud is evident in Figure 2 (upper); for overcast cloud cases, the mean cloud optical depth in 1995 (a drought year) was smaller than that of 1996.

An operational algorithm developed by Ineichen and Perez (1999) was used to derive a transmittance index (K) from the GOES-10 satellite for all of 1998. Figure 3 shows the comparison of the monthly mean transmittance index (K index) between Harvard Forest and Albany. Both sites show similar characteristic of aerosol seasonal variation, which is consistent with surface observation (shown in Fig. 1). However, under cloudy conditions the K index at the Harvard Forest is smaller before May than that at the Albany site, indicating thicker clouds over Harvard Forest than over Albany.

The spatial statistics of clouds, using the K index images from GOES to identify the cloud cluster areas over both sites, are shown in Figure 4. The left plot shows the statistics for all clouds, illustrating that large cloud systems with cloud cluster areas over 10^6 km² have the same statistics over both sites, due to large-scale circulation. On the other hand, small-scale cloud systems are markedly different due to local circulations. For cloud cluster areas less than 1000 km², shown in middle and right plots, the statistics show reversed characteristics for thick and thin cloud cases over two sites.

Figure 5 shows the relationship between the atmospheric transmittance for the Harvard Forest pixel (GOES-10, 1998) and the CO₂ flux; this figure can be compared to Figure 2 of Moore and Fitzjarrald (this volume), which is based on surface measurements of atmospheric transmittance.

Reference: Ineichen P. and R. Perez. 2000. Derivation of cloud index from goestationary satellites and application to the production of solar irradiance and daylight illuminance data, Theo. & Appl. Clim. (In Press).

---

Effect of Diffuse Light on Canopy CO₂ Uptake
K. Moore and D. Fitzjarrald

Freedman et al. (2000) have recently shown that the presence of boundary layer clouds is associated with enhanced CO₂ uptake at the midlatitude deciduous Harvard Forest. While we noted a similar association in a boreal jack pine forest (Moore et al. 2000), in that case decreased C uptake on clear days was accompanied by water stress and stomatal closure, an effect generally not observed at Harvard Forest. This study explores the relationships among clouds, diffuse light, and net photosynthesis at Harvard Forest.

A multifilter rotating shadowband radiometer (MFRSR) was used to obtain time series of direct, diffuse, total and reflected radiation at Harvard Forest in central MA during the growing season of 1998. Diffuse fraction (fraction of total solar radiation represented by the diffuse component) was related to the clearness index or atmospheric transmittance (S/S₀) by an equation similar to those published from other locations (Fig. 1). Midday hourly eddy covariance CO₂ flux is optimal at atmospheric transmittances of 0.5 to 0.6,corresponding to diffuse fractions of 0.6 to 0.7 (Fig. 2). We have found that canopy light use efficiency (mole CO₂ uptake/mole photons) is linearly related to diffuse fraction; light use efficiency under diffuse fraction of 0.6 is twice that under clear skies (