You are here

Harvard Forest Data Archive

HF084

Impact of Hemlock Woolly Adelgid on Canopy Throughfall in Southern New England 2002

Related Publications

Data

Overview

  • Lead: David Orwig, Bernhard Stadler
  • Investigators: Richard Cobb, Thomas Muller
  • Contact: David Orwig
  • Start date: 2002
  • End date: 2002
  • Status: completed
  • Location: South Central Connecticut, Harvard Forest
  • Latitude: +41.40 to +42.55
  • Longitude: -72.40 to -72.16
  • Elevation:
  • Taxa: Adelges tsugae (hemlock woolly adelgid), Betula lenta (black birch), Tsuga canadensis (eastern hemlock)
  • Release date: 2006
  • Revisions:
  • EML file: knb-lter-hfr.84.18
  • DOI: digital object identifier
  • Related links:
  • Study type: short-term measurement
  • Research topic: invasive plants, pests and pathogens
  • LTER core area: inorganic nutrients, disturbance
  • Keywords: carbon, hemlock, hemlock woolly adelgid, litter, nitrogen, precipitation, throughfall
  • Abstract:

    Exotic insect pests may strongly disrupt forest ecosystems and trigger major shifts on nutrient cycling, structure and composition. Although the immediate impact of these pests are frequently examined as physical disturbances (i.e., defoliation, decline in leaf area, and tree mortality) that initiate changes in ecosystem function, the insects often generate fundamental biochemical and trophic changes in tree canopies that may be equally important in altering ecosystem dynamics. Consequently, investigation of the linkages between canopy-level, ecosystem and environmental impacts may be critical for a thorough understanding of functional, structural, and compositional changes resulting from pest infestation. We sought to establish a better understanding of these linkages for the hemlock woolly adelgid (HWA), which is devastating hemlock forests in an expanding region across eastern North American and has the potential to eliminate this long-lived and extremely shade-tolerant species across much of its range. We examined the impact of the adelgid on hemlock needle chemistry and epiphytic microorganisms, litter production, and shoot growth in stands differing in their levels of infestation and linked these to shifts in canopy nutrient cycling and stand and landscape dynamics.

    HWA initiated major changes in canopy biomass and distribution. Whereas uninfested trees exhibit a decline in canopy biomass from the center to the periphery and a positive correlation between total needle litter and estimated biomass, infested trees support predominantly woody biomass, have significantly less total canopy biomass, produce less new foliage and exhibit no correlation between litter and canopy biomass. Foliar %N was strongly influenced by needle age and the level of infestation and was highest in young foliage supporting the highest densities of HWA. Foliar %C was unaffected by HWA or foliar age. Epiphytic microorganisms on hemlock needles exhibited little variation in abundance within canopies, but bacteria, yeasts and filamentous fungi were significantly more abundant on medium and heavily infested than uninfested trees.

    Throughfall chemistry, quantity, and spatial pattern were strongly altered by HWA. Beneath uninfested trees throughfall exhibits a strong gradient, decreasing in volumes from the canopy periphery to the trunk. Beneath infested trees the amount of throughfall is greatly increased, spatially unpatterned and characterized by higher concentrations of nitrogen compounds, carbon and cations.

    Across the southern New England landscape there is a strong south to north gradient of decreasing tree and sapling mortality and understory compositional change that correspond to the duration of infestation. Regionally, black birch, a nitrogen demanding species that is competitively enhanced by increasing nitrogen availability is profiting most from hemlock decline.

  • Methods:

    Overview

    At each site three throughfall samplers were placed beneath each tree, one close to the trunk, one at the periphery and one midway, to capture throughfall across the canopy spatial gradient. Each throughfall sampler covered 162.86 cm2 and remained in place through the entire experiment. Throughfall was collected every two weeks, from May 20th to August 31st. To determine the nitrogen content of needles on infested/uninfested shoots and egg mass biomass, needles and egg masses were detached with tweezers and composited for each shoot and determined with a Fisons dry combustion CHN autoanalyzer. At three sampling dates (May 22, July 1, and August 31, 2002) shoots from eastern hemlock trees were sampled for microbial analysis. Four shoots with needle age classes ranging from 0-3 years were cut with sterile scissors, transferred into a sterile Stomacher bag, stored in a cooler, and transported to the laboratory where they were frozen until microbial analyses could be completed.

    Experimental sites

    Infested sites included Devil’s Hopyard State Park (DH) and Selden Neck Preserve (SN), located in south-central Connecticut, while the uninfested reference site was the Prospect Hill (PH) tract of the Harvard Forest, located in north central Massachusetts. Average HWA infestation on shoots of the previous year was 5-10 egg masses at DH (in the following classified as medium infested site) and 15-25 egg masses at SN (heavy infested site) at the beginning of the experiment.

    Canopy biomass and shoot growth

    At each of the three sites, four trees were chosen that were approximately 6 meters tall and not overtopped by other trees. To estimate biomass above a throughfall sampler (for details see below), a digital photographic image was taken above each funnel at the beginning of the experiment. The central 30 percent of the image was converted to black and white pixels and analyzed using the Image Tool program (University of Texas Health Science Center at San Antonio, Texas; available at ftp://maxrad6.uthscsa.edu). The number of black pixel provides a standardized (relative) measure of tree canopy biomass above each throughfall sampler. A constant threshold value of 200 was used in the program to convert the images into black and white pixels. Note, such a value is arbitrary for setting the conversion algorithm. However, because we were interested in a relative comparison rather than absolute estimates of canopy biomass in relation to different degrees of infestation, applying the same procedure to every image provides a useful basis for comparing the effects of HWA on canopy biomass. We used only the central part of an image, rather than spherical canopy pictures because we seek to estimate the biomass of those parts of the crown, which most likely affects a particular throughfall sampler. This method is likely to overestimate the absolute biomass of infested canopies, because multiple layers of needles are poorly distinguished. However, a broad index is appropriate for the processes addressed in this study, and other approaches like the diffuse light penetration method are generally no more accurate or reliable (Lovett et al. 1996) because a number of variables like cloud cover, vertical position of the digital device or conversion algorithms can profoundly affect the results. Growth of newly developing shoots was estimated by measuring the length of 20 second-order shoots on each of four trees at each site. Growth was measured at breast height in the periphery of the trees following the aspect to which the samplers were pointing. Measurements were made biweekly at the same time, when throughfall was sampled.

    Sampling of hemlock shoots to determine microbial and HWA abundance and foliar N

    At three sampling dates (May 22, July 1, and August 31, 2002) shoots were sampled for microbial analysis. These dates cover the different activity periods within the life cycle of HWA: the growth of the first generation from early spring to early summer, the summer aestivation, and the growth of the second generation following the termination of diapause. Four shoots with needle age classes ranging from 0-3 years were cut with sterile scissors, transferred into a sterile Stomacher bag, stored in a cooler, and transported to the laboratory where they were frozen until microbial analyses could be completed. The sampled shoots were taken from the lower canopy offset from the throughfall samplers (i.e., close to the trunk, close to the canopy periphery and in between these locations). Shoots and canopy biomass directly above the samplers were kept intact. To determine HWA infestation intensity and C and N content of needles an additional three branches were cut from the periphery of each tree during sampling for foliar microbes and were returned to the laboratory on ice. Needles from six randomly selected shoots were detached and processed for C and N analysis and the number of HWA egg masses on the previous year’s shoot was counted.

    To get a rough estimate on the total number of egg masses in the canopy of medium and heavily infested trees we counted the number of one-year-old shoots on an entire branch near the base of the tree, multiplied this number by the number of branches on the tree, and then multiplied this number by the average number of egg masses per shoot. Branches at the base develop most new shoots mainly at the periphery and the addition of new growth appears to be similar on branches higher up in the canopy. In addition, we performed only relative comparisons between differently infested trees.

    Microbiological analyses (data not available)

    Needles of hemlock were carefully dislodged from the twigs. Three grams of each sample of green needles (all age-classes pooled) from the twigs were blended for 2 minutes in 145 ml sterile distilled water using a Stomacher lab blender. Leaf washings were logarithmically diluted in 1/4 strength Ringer’s solution and spread-plated onto 1/10 strength Tryptic soy agar (Merck; pH 7.2), supplemented with 0.4g l-1 Cycloheximide (Merck), to enumerate the numbers of aerobic heterotrophic bacteria in the sample. Yeasts and filamentous fungi were grown on a Sabouraud-1% dextrose-1% maltose agar (Merck; pH 5.5), to which 0.4g l-1 Chloramphenicol (Berlin-Chemie) was added to suppress bacterial growth. All plates were incubated at 25 °C for 5 days. Results were expressed as colony forming units (CFU) per gram of needle dry matter.

    Throughfall sampling and needle litter collection

    At each site three throughfall samplers were placed beneath each tree, one close to the trunk, one at the periphery and one midway, to capture throughfall across the canopy spatial gradient. Each throughfall sampler covered 162.86 cm2 and remained in place through the entire experiment. Samplers usually faced to the southwest. The neck of the throughfall funnels was plugged with a nylon filter to prevent the entry of debris and insects, and bottles were wrapped in aluminum foil to protect against light and heat. At each sampling date the bottles and funnel plugs were replaced with clean ones and the number of hemlock needles in the funnels were counted. Throughfall was collected every two weeks, from May 20th to August 31st. In early July the sampling period was three weeks due to low amounts of precipitation. A single rain sampler was installed at each site to collect bulk precipitation for chemical analysis. Precipitation data for each area were also obtained from meteorological stations at Harvard Forest, MA, New Haven, CT and Middletown, CT.

    Chemical analyses of throughfall, nitrogen content of needles and egg masses

    All throughfall samples were immediately filtered through a 0.45-mm cellulose acetate membrane and frozen until analysis. Dissolved organic carbon (DOC) was determined as CO2 after persulphate-UV-oxidation (Foss Heraeus, Liquid TOC). Ammonium- N (NH4 -N) and nitrate-N (NO3 -N) were measured by ion chromatography (Dionex, Idstein, 2000i-SP). Dissolved organic nitrogen (DON) was calculated from the following relationship: DON = Ntotal - ( NH4 -N + NO3 -N). Total nitrogen (Ntotal ) was measured as NOx after thermo-oxidation at 700°C (Abimed: TN-05). To determine the nitrogen content of needles on infested/uninfested shoots and egg mass biomass, needles and egg masses were detached with tweezers and composited for each shoot. In total, 85 needle samples and 32 egg mass samples were analyzed. Foliage and egg masses were dried at 50o C for 48 hours. Foliar material was then ground to pass through a 20-µm screen. Egg sack mass was determined with a fine scale balance accurate to 0.001 mg. Total C and N content was determined with a Fisons dry combustion CHN autoanalyzer (Milano, Italy). Approximately 20% of foliar samples were analyzed in duplicate.

    Change in stand structure due to HWA infestation

    As part of a large study examining landscape patterns of hemlock decline in Connecticut (Orwig et al. 2002), 114 hemlock stands were quantitatively sampled for intensity of HWA infestation and stand and site characteristics. Stands were delineated from 1:80000 scale black and white photographs and were randomly selected across the study region. Overstory and understory vegetation was sampled in one fixed-area (20 x 20 m) plot in a representative portion of each stand. All trees (stems ³ 8 cm diameter breast height (dbh)) were tallied by species and dbh, and assigned a canopy position based on a visual estimation of the amount of intercepted light received by the tree crown (Smith, 1986). Hemlocks that died within the previous 2 to 4 years, which was determined by extensive retention of fine twigs in the crown, were tallied to determine species composition prior to HWA infestation. All saplings (less than 8 cm dbh and greater than 1.4 m tall) were tallied by species. The percent cover of each herb, shrub, and woody seedling species was estimated in each plot using a modified Braun-Blanquet scale.

    Data analyses

    Throughfall fluxes were converted to mg m-2 14 days-1 with the exception of the one sampling period comprising 21 days. However, because the sites received different amounts of rainfall it is not useful to compare total fluxes between sites with respect to specific sampling dates. Rather we restricted the analysis of fluxes to the variability within the canopies of trees showing different levels of infestation. We used regression analysis to investigate the relationship between the amount of throughfall and the concentrations of specific compounds. In addition, we compared the results of the reference sites with those of the infested sites using GLM ANCOVA, with rain volume and infestation as main effects and the index of biomass as a covariable, testing for significant interactions between volume and infestation. The analysis was restricted to throughfall lower than 40 mm per sampling period to account only for the linear part of the relationship between these two variables. Differences in CFU’s of microorganisms were tested with one-way ANOVAs using Bonferroni correction.

    Foliar chemistry was analyzed with a two-way repeated measures ANOVA using infestation and needle age as the main effects. To determine the effects of infestation for any single foliar sampling we employed a one-way ANOVA with site (infestation level) as the main effect. Recovery of two-year-old needles was inconsistent on infested trees resulting in poor replication for this age class and exclusion from statistical analysis. Repeated measures ANOVAs were also used to test for differences in the spatial variability of litterfall, amount of throughfall and throughfall fluxes within trees. For the analyses the normality assumptions of the data were checked and log x+1 transformed when necessary to normalize variances across treatments.

    To assess the relative importance of environmental and stand variables in controlling understory vegetation dynamics, we utilized Mantel tests (Mantel, 1967; Manly, 1997a), which includes space (i.e., geographic location) as a predictor variable in the analysis. This technique performs a linear regression on distance matrices generated from dependent variables (birch sapling density, birch seedling cover, total understory species richness and herbaceous species richness) and predictor variables (space, latitude, slope, aspect, elevation, and stand size). Prior to analysis, aspect values were transformed from circular variables to a measure relevant to vegetation as: aspect = cosine (45 - azimuth degrees) + 1 (Beers et al., 1966). Values range from 0 on southwestern slopes commonly exposed to the sun to 2 on the least exposed northeastern slopes. In addition, latitude and longitude were converted to distance measures with the same units (i.e., km from the equator or prime meridian, respectively).

  • Use:

    This dataset is released to the public under Creative Commons license CC BY (Attribution). Please keep the designated contact person informed of any plans to use the dataset. Consultation or collaboration with the original investigators is strongly encouraged. Publications and data products that make use of the dataset must include proper acknowledgement.

  • Citation:

    Orwig D, Stadler B. 2006. Impact of Hemlock Woolly Adelgid on Canopy Throughfall in Southern New England 2002. Harvard Forest Data Archive: HF084.

Detailed Metadata

hf084-01: throughfall

  1. date: sampling date of throughfall, needles, foliage etc.
  2. site: location of sampling
    • DH: Devil's Hopyard, CT (medium infested)
    • PH: Prospect Hill, Harvard Forest, MA (control)
    • SN: Selden Neck, CT (heavily infested)
  3. sample: identification of tree within site (n = 3 per site)
  4. black.per: indicator of overhead biomass determined from percent of black pixels in canopy photographs (unit: number / missing value: NA)
  5. position: sampling location under canopy
    • 1: trunk
    • 2: middle canopy
    • 3: periphery
  6. volume: volume of throughfall collected in sampler (unit: milliliter / missing value: NA)
  7. shoots: number of one-year shoots collected above each collector (unit: number / missing value: NA)
  8. eggs: number of egg masses per shoot above each collector (unit: number / missing value: NA)
  9. needles.s: number of needles on one-year shoots above collectors (unit: number / missing value: NA)
  10. needles.f: number of needles collected in funnel at each sampling (unit: number / missing value: NA)
  11. no3.n: nitrate N collected in funnel (unit: milligramsPerLiter / missing value: NA)
  12. nh4.n: ammonium N collected in funnel (unit: milligramsPerLiter / missing value: NA)
  13. nt: total Nitrogen collected in funnel (unit: milligramPerMeterSquaredPerDay / missing value: NA)
  14. don: dissolved organic nitrogen (Nt-NH4-NO3) in funnel (mg N m-2 d-1) (unit: milligramPerMeterSquaredPerDay / missing value: NA)
  15. toc: total organic carbon in funnel (mg C m-2 d-1) (unit: milligramPerMeterSquaredPerDay / missing value: NA)
  16. ca2: calcium collected in funnel (unit: milligramsPerLiter / missing value: NA)
  17. k: potassium collected in funnel (unit: milligramsPerLiter / missing value: NA)
  18. mg: magnesium collected in funnel (unit: milligramsPerLiter / missing value: NA)
  19. mn: manganese collected in funnel (unit: milligramsPerLiter / missing value: NA)
  20. na: sodium collected in funnel (unit: milligramsPerLiter / missing value: NA)
  21. cl: chlorine collected in funnel (unit: milligramsPerLiter / missing value: NA)
  22. so4: sulphate collected in funnel (unit: milligramsPerLiter / missing value: NA)
  23. s: sulphur collected in funnel (unit: milligramsPerLiter / missing value: NA)
  24. growth: average shoot growth above samplers (unit: millimeter / missing value: NA)

hf084-02: precipitation

  1. period: number of collection period (1-7)
  2. start.date: start date of collection period
  3. end.date: end date of collection period
  4. hf.prec: total precipitation for sampling period at Harvard Forest, MA (unit: millimeter / missing value: NA)
  5. nh.prec: total precipitation for sampling period at New Haven, CT (unit: millimeter / missing value: NA)
  6. md.prec: total precipitation for sampling period at Middletown, CT (unit: millimeter / missing value: NA)
  7. hf.trunk: total throughfall amounts for site and canopy position (unit: millimeter / missing value: NA)
  8. hf.middle: total throughfall amounts for site and canopy position (unit: millimeter / missing value: NA)
  9. hf.periphery: total throughfall amounts for site and canopy position (unit: millimeter / missing value: NA)
  10. dh.trunk: total throughfall amounts for site and canopy position (unit: millimeter / missing value: NA)
  11. dh.middle: total throughfall amounts for site and canopy position (unit: millimeter / missing value: NA)
  12. dh.periphery: total throughfall amounts for site and canopy position (unit: millimeter / missing value: NA)
  13. sn.trunk: total throughfall amounts for site and canopy position (unit: millimeter / missing value: NA)
  14. sn.middle: total throughfall amounts for site and canopy position (unit: millimeter / missing value: NA)
  15. sn.periphery: total throughfall amounts for site and canopy position (unit: millimeter / missing value: NA)

hf084-03: flux

  1. flux: measured flux
    • DOC: dissolved organic carbon
    • DON: dissolved organic nitrogen
    • Nt: total nitrogen
  2. site: site code
    • DH: Devil's Hopyard, CT (medium infested)
    • PH: Prospect Hill, Harvard Forest, MA (control)
    • SN: Selden Neck, CT (heavily infested)
  3. position: canopy position
  4. period1: total fluxes (mg m-2 15 weeks-1) for each sampling period, site, and canopy position. See precipitation metadata for actual dates. (unit: milligramsPerSquareMeter / missing value: NA)
  5. period2: total fluxes (mg m-2 15 weeks-1) for each sampling period, site, and canopy position. See precipitation metadata for actual dates. (unit: milligramsPerSquareMeter / missing value: NA)
  6. period3: total fluxes (mg m-2 15 weeks-1) for each sampling period, site, and canopy position. See precipitation metadata for actual dates. (unit: milligramsPerSquareMeter / missing value: NA)
  7. period4: total fluxes (mg m-2 15 weeks-1) for each sampling period, site, and canopy position. See precipitation metadata for actual dates. (unit: milligramsPerSquareMeter / missing value: NA)
  8. period5: total fluxes (mg m-2 15 weeks-1) for each sampling period, site, and canopy position. See precipitation metadata for actual dates. (unit: milligramsPerSquareMeter / missing value: NA)
  9. period6: total fluxes (mg m-2 15 weeks-1) for each sampling period, site, and canopy position. See precipitation metadata for actual dates. (unit: milligramsPerSquareMeter / missing value: NA)
  10. period7: total fluxes (mg m-2 15 weeks-1) for each sampling period, site, and canopy position. See precipitation metadata for actual dates. (unit: milligramsPerSquareMeter / missing value: NA)
  11. total.flux: total flux for all periods (mg m-2 15 weeks-1) for each site and canopy position (unit: milligramsPerSquareMeter / missing value: NA)