EXPERIMENTAL DESIGN AND METHODS: (B.Kaye, Updated 2010 V.Salmon)

SITES AND VEGETATION TYPE CLASSIFICATION:

Toolik 2009
Vegetation types are moist acidic tussock, dry heath, and wet sedge.

Barrow 2009
All vegetation is classified as wet coastal tundra.
Zackenberg 2006
Vegetation types are based on location and species composition and are classified as: Cassiope heath, Vaccinium heath, Dryas heath, Salix snow bed, Fen, Grassland or Abrasion plateau according to the descriptions and mapping by Bay 1998.

Abisko 2005
Veg types are: Betula, Heath, Rocky, Salix, Sedge and Rocky. Vegetation types were determined by dominant species in the plots. The vegetation type 'rocky' was dominated by bare ground, cryptogamic crust and a mix of prostate shrubs and forbs.

Svalbard 2005
Veg types are: Salix polaris, Mixed heath, Dryas, Equisetum, Dry Heath, Graminoid, Cassiope, Salt Marsh, and Polar desert. Vegetation types were determined by the dominant species in each harvest or by plot characteristics (soil, nutrients, disturbance regime, exposure, etc.). If one species or functional type (I.E. graminoids) dominated a site (>50%) it was classified as that type. If no vegetation type was dominant the harvest was classified as mixed.

Some sites were classified by characteristics. For example, the 'salt marsh' vegetation type was classified that due to it's unique characteristics (soil, nutrients, disturbance regime, etc.) even though it was dominated by graminoids. Dry heath was also separated from mixed heath: dry heath sites had extremely low vascular vegetation coverage (riverbed sites and others, dominated by non-vascular crust and bare ground). Polar desert was designated for harvests taken on top of Berget Plateau.

Abisko 2004
Veg types are: Betula, Heath, Rocky, Salix, Dryas, Wet meadow, Wet sedge, Mesic Meadow, Snow bed, and Tussock tundra. Vegetation types were determined by the dominant species in each harvest or by plot characteristics. See '2003-2005 All Summary Data' for a more in depth description of flux plot characteristics (same name as harvest IDs). Heath sites consisted of mixed vegetation. Rocky sites were in exposed areas with mix of prostate shrubs and forbs. Snowbed sites were relatively unproductive sites, covered by snow late into the growing season. Wet meadow sites had mixed species, wet sedge sites were dominated by sedges. Mesic meadow sites were species rich.

Toolik 2004
Veg types are: Wet sedge, Dry Heath, Salix, Betula, Moist-Acidic tundra, Non-acidic tundra, and Non-acidic non-tussock tundra. All but two harvests (Salix and Betula taken at Imnavait creek) were from fertilized areas of the Toolik LTER long term fertilization experiment area.

Toolik 2003
Veg types are: Arctostaphylos/dryas, Dryas, Dryas/Carex, Carex, Wet sedge, Betula/VVI, Betula, Betula dry, Betula water track, Salix, Salix water track, Dryas/dwarf fireweed, Ledum/empetrum/vvi, Hilltop tussock (dry), Tussock tundra (wet), and Open tussock. Vegetation types were determine by dominant species in the harvest or by plot characteristics (I.E. tussock sites.)

HARVEST METHODS:
Methods varied slightly from year to year. Methods for each year are outlined.

2003 Toolik
30 - 20x20 cm quadrats were harvested in a range of vegetation types around Toolik Lake and the Imnavait Creek watershed. Prior to harvesting, SKYE sensor readings (see explanation below) were taken in order to calculate NDVI of the harvest area. Measurements were also taken using a LAI 2000; however this instrument proved to be ineffectual at estimating leaf area in tundra vegetation. All above ground vascular plant vegetation was collected from the harvest. Vegetation was sorted into species type and was scanned and analyzed for leaf area using WinRhizo software. Samples were dried and weighed, and then discarded.

For C-N analysis, samples of the three most dominant species from each flux plot were collected on the 8th of August. See file 2003-2004gsfluxleafN.dat Leaf samples for C:N analysis from 2003 were processed in Fall 2003 by Lorna Street at the University of Edinburgh, UK. Samples were ground and analyzed for carbon and nitrogen concentration using CE-440 Elemental Analyzer (North Chelmsford MA, USA).

2004 Toolik
18 - 20x20 cm quadrats were harvested using the same methods as in 2003. Harvests were taken from control and fertilized areas. After samples were dried and weighed samples were discarded.

For C-N analysis, samples of the three to four of the most dominant species from each flux plot were collected. C-N analysis was done in Woods Hole, Fall 2005. (Petioles were removed before grinding leaves for C-N analysis.) See file 2003-2004gsfluxleafN.dat

2004 Abisko
23 - 20x20 cm quadrats were harvested from the center of the fluxing plots at the Latnja and Stepps sites. Harvests were not taken from the Paddus site. In the SB1 and SS1 sites a 40x40 cm sample of the tall shrubs was harvested above the 20x20 quadrat. All biomass to the top of the moss layer was harvested and the moss was collected separately. Stems in the moss were considered belowground and were discarded or left in the ground. Harvests were sorted by species and by production fraction (new leaf, old leaf, new stem, old stem, inflorescence).

All leaves were scanned using the WinFolia systems and a small portable Canon scanner. The biomass was dried at 70 deg. C for at least 48 hours and was weighed. Skye sensor scans and Unispec Spectral Analysis scans (see explanation below) were not taken of the harvest specifically, but Unispec and/or Skye sensor scans were taken of the flux plots from which the harvests were taken.

New leaves and old leaves were ground and analyzed for C-N in Woods Hole, Fall 2005.

All petioles were clipped from leaves before they were ground. Samples were reweighed after petioles were removed so that N was not over-estimated. Where sub samples were taken (often only sub samples were sorted into production fractions) dry weight ratios were used to estimate the leaf biomass of the total sample. Several harvests were not sorted at all; only dry weights of biomass by species were taken. For these harvests production fractions from similar harvest types were used to estimate leaf area, leaf biomass and new/old stem biomass.

Abisko/Svalbard 2005
93 - 26 cm diameter harvests were taken. In Abisko in June, 28 harvests were taken in 5 different vegetation types (sedge, heath, rocky, betula, and salix) near fluxing plots. In Abisko in August, 15 harvests were taken, one from each flux plot. In Svalbard, 50 harvests were taken from fluxing sites (occasionally from within a fluxing plot). Before each harvest was taken, Unispec Spectral Analysis scans were taken and Skye sensor scans were taken (no Skye sensor scans for Abisko June 2005) in order to calculate NDVI. For some harvests, a Unispec scan was also taken post harvest, in order to estimate background reflectance, and reflectance of non-vascular vegetation. A visual estimate of moss cover in the plot was made after harvesting.

We bagged and refrigerated the harvested material until it could be sorted (within max. 2 days). In the lab we separated material into species, separated live leaves from stems and litter (removing petioles of all leaves, except for the June harvests, in which case petioles were removed once the leaf material was dried). While leaves were fresh, leaf area was measured using a Canoscan scanner connected to a laptop. The images were analyzed using Winfolia software . After scanning the leaf samples were dried at 65oC for a minimum of 48hrs and were then weighed.

We took a sub samples if the leaves from a species were too numerous to scan. Dry weight ratios were then used to estimate total leaf area of the sample. When samples were large, more than one sub sample was taken and an average of each individual leaf area/dry weight ratio was used to scale up to the leaf area of the total sample.

In addition to leaf area sub samples, in Abisko, some species were too numerous to pluck all the leaves from the stems (Empetrum nigrim and Lycopodium spp). In June Abisko harvests, sub samples were taken on the basis of fresh weight. Leaves from the sub samples were plucked from the stems, scanned, dried and weighed; the remainder of un-plucked material was discarded. Fresh weight ratios were used to estimate the dry weight and leaf area of the total sample. In Abisko in August we kept and dried the non-plucked remainder, the live leaves from the sub samples, and the dead leaves from the sub sample. We shook the dried leaves off of the stems of the remainder and used dry weight ratios to estimate leaf area and dry weight of live leaves in the total sample.

There were problems scanning empetrum and other small-leaved species due to shadows created by the leaves. Thick-leaved species also presented a problem by pushing the lid of the scanner higher and causing greater shadows. Varying the contrast of the scanned image caused large differences in leaf area. In August we were able to use a camera which takes an image of transmitted rather than reflected light which prevented the problem of shadows. Empetrum seemed to present the largest over-estimation of leaf area (due to it's small size and high abundance in most harvests) so we used the camera to determine empetrum leaf area. We also re-scanned empetrum samples from June, under the assumption that drying the empetrum leaves did not significantly change their area. Care had to be taken when scanning pale colored leaves. Leaves had to be turned so that the darker side of the leaf was being scanned, otherwise leaf area of highly pubescent, pale leaves would be underestimated.

Samples from 2004-2005 were ground and analyzed for CN between October 2005 and April 2006 by Brooke Kaye at the Ecosystems Center in Woods Hole, MA. Samples were briefly re-dried at 60oC (24 hours) and put in a dessicator. Smaller samples (less than ~0.3 grams, depending on leaf size) were ground for approximately one minute in the Wig-L-Bug in the grinding room at the Marine Biological Lab. Larger samples (larger than ~0.3 grams, depending on leaf size) were ground in the Retsch MM200 double grinder for approximately three minutes, or until samples were finely powdered with no remaining visible plant fibers. If leaves still had petioles attached I removed the petioles with scissors and discarded them before grinding the leaf sample. All ground samples were stored in plastic 20ml scint vials until C-N analysis could be done.

For Abisko 2005 harvests, all leaf samples from the harvests are stored in scint vials (species with low mass/leaf area were simply stored in the 20ml scint vials and not ground in order to keep all the species from a harvest together.) All C-N leaf samples from Toolik 2004 were ground and are stored in the 20 ml scint vials. For Abisko 2004 and Svalbard harvests, only species constituting > 85% of leaf area were ground and stored in 20ml scint vials. Remaining leaves are stored in their original envelopes.

For C-N analysis, tins were packed with 3-4 milligrams of powdered leaves. One tin was packed for every sample type. Every 10th sample had a replicate sample packed from it in order to check the reproducibility of values and the uniformity of the ground sample

After prep samples were run, leaves were analyzed in auto-run mode (non-filter type samples). After every 10 samples a blank tin and a K-factor (standard - Acetanilide) were run to check accuracy of values. Samples were run on a Perkins Elmer CHNS/O Series II analyzer, owned by the Ecosystems Center and maintained by Don Burnette in 2005/2006.

Zackenberg 2006
30 26cm diameter harvests, 25 30cm x 30cm harvests (of 30cm x 30cm flux plots) and 30 20cm diameter harvests (of moss NDVI study plots) were taken. Methods for Unispec measurements are identical to 2005 except the black foreoptic cable was used for all measurements; the height of the measurement adjusted such that the field of view matched the size of the plot as closely as possible. Tetracam images were also taken before and after each harvest, NDVI data from those images is not yet available.

For the 26cm diameter and 30cm x 30cm plots, moss cover was estimated visually after harvesting. For the 20cm diameter plots, moss cover was estimated using a 2cm x 2cm grid placed over the circle. The presence or absence of moss in each square was then recorded on a map of the plot (a square could also have 50% moss). Total cover was then counted from the maps. 1 or 2 leaves of up to 3 species within each 20cm circle plot were removed before the rest of the harvest in order to take leaf level Unispec scans. The leaf area and dry weight of these leaves is included in the harvest data by species and in the total leaf area and biomass data for each plot.

Harvest and lab methods were the same as 2005; dry weight ratios were used to calculate total leaf area of samples from plucked sub samples. As there was no Empetrum it was possible to pluck all leaf material for all samples. All petioles were removed.

CN analysis completed by J. Peters in 2009 on a Perkins Elmer CHNS/O Series II Analyzer, owned by The Ecosystems Center and maintained by Don Burnette. For samples with no direct CN analysis (i.e., sub samples collected for leaf level NDVI scan) %C and %N values were assigned as follows: Within each harvest plot, CN values from one analyzed sample were assigned to all the unanalyzed samples of the same species. If a species did not have any samples that were directly analyzed, CN values were assigned from a species that was collected in the same harvest and represented the same functional group. Averages were used if there were multiple species in the same functional group as the unanalyzed species.

Toolik/Barrow 2009
A total of 77 circular (20 cm diameter) harvests were taken. At Toolik Lake, harvests were taken on four days over the course of the summer : June 24th, July 13th, July 30th and August 14th. For each harvesting date, 5 harvests were taken in three different vegetation types: Moist acidic tundra, Wet sedge and Dry heath . The exception was the August 14th harvest date at which point only three harvests were taken per vegetation type. Harvest locations were chosen to encompass a range of leaf areas and vegetation within each vegetation type. At Barrow, harvests were taken on three days (July 29th, July 30th, and August 3rd) in a survey of dominate ground cover.

For all plots, Unispec Spectral Analysis scans were taken prior to harvesting. Harvested material was bagged and refrigerated until sorting (max. 2 days delay). A visual estimate of moss cover in the plot was made after harvesting. In the lab, material was separated by species, live leaves were separated from stems and litter, and petioles were removed from Betula leaves. While leaves were fresh, leaf area was measured for each species. For the June and July harvests at Toolik, leaf area was measured using a Licor portable leaf area meter. For the August harvests at Toolik and all Barrow harvests, leaf area was measured using a Canoscan scanner connected to a laptop. The images were analyzed using Winfolia software. After scanning, all leaf samples were dried at 65oC for a minimum of 48hrs and were then weighed.

CN analysis was completed by V. Salmon in January 2010 on a PE 2400 Series II CHNS/O Elemental Analyzer owned by The Ecosystems Center and maintained by Marshall Otter. Samples were re-dried and stored in a dessicator. Larger samples were ground using a Retsch MM200 double grinder while smaller samples were ground using a Wig-L-Bug. Some samples were too small to grind (1-2 milligrams) and were left whole. All samples were stored in plastic 20 ml scintillation vials. Depending the amount of available sample, between ~3-5 milligrams of homogenized material was packed into tins for analysis. For the small, unground samples, all available material was packed into a tin for analysis. Every tenth sample was replicated and a Quality Assurance sample of apple leaf (known %c and %N) was run for every twenty unknown samples.

NDVI MEASUREMENT

The unispec spectral analyzer measures reflected light intensity in 256 portions of the visible spectrum from ~300nm to ~1100nm. A foreoptic cable is connected to the instrument and directed at the target, a scan is then taken over a period of milliseconds.

The end of the fibre optic was centered above the harvest and held at a height above the ground surface that resulted in a view of the vegetation approximately the size of the harvest frame. Scans were taken with the fiber optic held higher if shrub canopy was tall (some Salix pluchra plots in Abisko 2005). Incident radiation was measured using a reference standard so that reflectance could be calculated as a percentage of incoming solar radiation.

The Skye portable NDVI sensor measures radiation reflected from the vegetation (without incident radiation) in the wavebands 570nm-680nm and 725nm - 1100nm. NDVI is then calculated as below. The Skye sensor was centered above the harvest and held at a height that resulted in a view of the vegetation approximately the size of the harvest frame.

NDVI = (RII-RI)
(RI + RII)

where RI = average reflectance from 570nm to 680nm
RII = average reflectance from 725nm to 1000nm.

CALCULATIONS:

Abisko 2004 and Abisko/Svalbard 2005:
A minimum set of species that accounted for >85% of total leaf area for each harvest were selected and analyzed for C:N (Van Wijk et al. 2005). For species that were not analyzed for C:N from each harvest (because these species were low in biomass for that particular harvest) estimates of total C and N for that species were made based on average C and N percentages for that species for that approximate harvest date (Grouped by Early June, Late June, and August for Abisko 2005. Grouped by site, Latnja or Paddus, for Abisko 2004. All Svalbard harvests were grouped together.) A select number of additional samples were run so that no species weighing > 0.1g was left without a C:N estimate. Total C and N of all species was then summed for each 20cm x 20cm harvest. For leaf area, biomass and N by species see file 2003-2005gsharvest.dat

LAI = leaf area/ harvest area

Biomass per area = dry biomass/harvest area

Harvest SLA = leaf area / (dry biomass/1000)

Fc per ground = foliar C/harvest area

Fn per ground = foliar N/harvest area

Fn per leaf area = foliar N/leaf area

Fn per biomass = foliar N/(dry biomass/1000)

Harvest area = 0.040m2 (2003-2004) or 0.053m2 (2005)

Total N and C are not given for harvests from 2003-04 because CHN analysis was not performed on the harvest leaf material (the only leaf samples taken for CHN were from the flux plots themselves). For the results of CHN analysis on leaf samples taken from the ITEX flux plots see the file 2003-2004gsfluxleafN.dat

COMMENTS:

FOR MORE INFORMATION CONTACT: Gus Shaver, The Ecosystems Center, Woods Hole, MA, 02543, USA

REFERENCE CITATIONS:
Bay, C. (1998) Vegetation mapping of Zackenberg Valley Northeast Greenland. Danish Polar Center.

Shaver, G.R., Street, L.E., Rastetter, E.B., van Wijk, M.T., Williams, M. (2007) Functional convergence in regulation of net CO2 flux in heterogeneous tundra landscapes in Alaska and Sweden. Journal of Ecology 95:802-817.

Street, L.E., Shaver, G.R., Williams, M.van Wijk, M.T. (2007) What is the relationship between changes in canopy leaf area and changes in photosynthetic CO2 flux in arctic ecosystems? Journal of Ecology 95: 139-150.

Van Wijk, M. T., Williams, M., Shaver, G.R. (2005) Tight coupling between leaf area index and foliage N content in arctic plant communities. Ecosystem Ecology 142: 421-427.

Williams, M., Street, L.E., van Wijk, M.T., Shaver, G.R. (2006) Identifying differences in carbon exchange among arctic ecosystems. Ecosystems 9: 288-304.

See Methods