1. Site description-
a. Site type: The site is an experimental forest located in the White
Mountains of New Hampshire, operated under the supervision of the
Northern Research Station of the United States Forest Service.
b. Geography: The Hubbard Brook Experimental Forest (HBEF) is
approximately 32 km2 and ranges in altitude from 222 to 1,015 m.
“The HBEF is located within the townships of North Woodstock (~60 %),
Ellsworth (~40 %), and Warren (<1 %), NH, within the White Mountain
National Forest of north central New Hampshire. Coordinates of 43°56′
N, 71°45′W bisect the area. The Atlantic Ocean is about 116 km to the
southeast. The nearest town is West Thornton, NH.” (Likens 2013: 5)
c. Habitat: The ecosystem is a mixed temperate hardwood forest
dominated by sugar maple (Acer saccharum), American beech (Fagus
grandifolia) and yellow birch (Betula alleghaniensis).
d. Geology: Bedrock geology is composed of quartz-mica schists and
quartzite of the Silurian Rangeley Formation. Significant glacial
deposits overlay much of the landscape. Soils are of the spodosol and
inceptosol orders.
e. Watersheds/hydrology:
“The experimental watershed-ecosystems range in size from 12 to 77 ha
and in altitude from 500 to 910 m. These headwater watersheds are all
steep (average slope of 20–30 %), with well-incised channels and
relatively distinct topographic divides. The height of the land
surrounding each watershed-ecosystems, the area, and the topography
all have been determined from ground surveys and aerial photography
and most recently augmented by Laser Imaging Detection and Ranging
(LIDAR). Experimental Watersheds 1–6 face S to SE; Watersheds 7–9 face
N to NE.”
(Likens 2013: 7)
f. Site history: Secondary forest, logged ca. 1910. Extensive
treatment of the site history can be found in Likens and Bormann
(1979), Holmes and Likens (2016) and on the HBES web page at
http://www.hubbardbrook.org/overview/sitedescription.shtml
g. Climate: mean annual precipitation - 1434 mm, mean annual
temperature - 6°C
2. Sampling design: The sampling design is exhaustively described in
the USFS technical report by Buso et al. (2000).
Executive Summary of Methods: Water samples are collected every Monday
from all stream sites and rain gauges. Water is collected in acid
washed, HDPE bottles and returned to the lab for filtration. At the
time of collection, the field technician records gauge height for each
watershed and precipitation volume for each rain gauge. Stream samples
are always collected from the same location.
Samples are returned to the USFS Pierce laboratory at Hubbard Brook
within hours of collection. There samples are analyzed for DIC, pH,
electrical conductivity. Beginning in 2017 they were also examined
spectrophotometrically at five wavelengths. A subsample is filtered
and shipped to the USFS Durham laboratory for subsequent analyses. The
remainder of the (unfiltered) sample is stored in the water sample
archive at USFS Pierce lab.
Details on the Precipitation Chemistry Samples: Perhaps the most
important aspect to explain here is the way in which the precipitation
dataset is curated. Contamination issues always plague rain sampling
efforts, and the policy for the HBEF long-term record has been to
curate the dataset heavily so that only high quality sample
information is retained. To help explain this we provide the following
historical overview of the approach to QA/QC of precipitation samples
Precipitation Chemistry records within the dataset were compiled from
separate sites at Hubbard Brook to represent the “best” sample (read
“cleanest”) for each weekly collection period. In the early years of
the HBES, the issue of precipitation contamination by coarse particles
(e.g., seeds and leaves) was clearly seen as a potential problem in
open collectors. While not exclusively a ‘growing season’ problem, the
worst incidents tended to occur in the summer and fall. Several
attempts were made to eliminate foreign objects, such as screens, wads
of poly-wool, and use of wet-only automated collectors, but these were
problematic in their own way. Screens had to be kept clean, and
sometimes precipitation was ‘filtered’ through trapped debris. It was
eventually decided that for ecosystem inputs, a bulk (continuously
open) collector was the most appropriate choice. This required the
positioning of several collectors so that at least one sample
collected from at least one site each week would be reasonably clean.
Although occasional grossly contaminated samples were not analyzed
(that is, thrown out after collection), most precipitation samples
were sent to the analytical lab for solute determination even in the
presence of amorphous fine particulates. It turned out that some fine
particulates seemed to dissolve and influence the analysis (probably
bird feces), while others appeared to be relatively inert (probably
fine pieces of bark or grit). There was no evidence, in the early days
of the record, that selecting chemistry from different sites around
HBEF would result in significant bias. In fact, the existence of
particulate contamination was considered a more important factor in
QA/QC than any potential difference in chemistry across the valley.
Long-term data since then bear this out as a general truth. Of course,
the resulting input flux values were always calculated using the
closest volumetric rain gauge, or a pattern of rain gauges in the case
of a whole watershed, since the variation in precipitation volume is
significant.
From 1963-1990 the cleanest estimate was selected to represent each
part of the valley. From 1990-2008 samples were averaged when more
than one was determined to be uncontaminated. In 2008, it became clear
that this averaging had no effect and the procedure reverted to the
original process of selecting a single rain gauge sample each week.
This approach was used until 2013.
After 2013, we report the raw data for all uncontaminated samples
collected from any rain gauge.
Research methods for sample analysis: The analytical methods prior to
2009 are exhaustively described in the USFS technical report by Buso
et al. (2000). Samples were analyzed at the Cary Institute of
Ecosystem Studies (Millbrook, NY)
As of June 2013, after careful inter-laboratory comparisons, chemical
analyses were transferred from the Cary Institute of Ecosystem Studies
to the USDA Forest Service Lab in Durham, NH.
Current Protocols (2013 to present):
Streamwater samples are collected weekly in watersheds 1-9 upstream of
the weirs, at the main stem of Hubbard Brook, and the M.Lake Outlet.
Dedicated, site-specific, pre-labeled collection bottles are
thoroughly DIW rinsed in between collection days and checked for
cleanliness (Sp.Cond ≤ 2uS/cm) prior to sample collection. During each
collection, a 1000 mL and a 60 mL LDPE bottle are filled and rinsed 3
times with streamwater prior to sample collection from the collection
site. Samples are collected with no headspace in the 60 mL bottle to
allow subsequent DIC analysis. A field temperature measurement is
taken with an Oakton Acorn series thermistor, and gage height readings
are recorded, along with any relevant field notes. One duplicate
sample is taken each week at a rotating site for QA/QC.
Prior to water year 2013 (June 1, 2013), samples were not filtered or
frozen prior to analysis. Since June 1, 2013 samples are brought back
to the Pierce Lab, and immediately vacuum filtered through 0.7um GF/F
filter into two pre-labeled 60 mL bottles. One bottle is frozen and
the other is refrigerated before being shipped to the USFS Durham lab
for cation/anion analysis.
DIC samples are immediately refrigerated and run within 24 hours of
collection on a Shimadzu 8A Gas Chromatograph. pH, ANC and Specific
Conductivity measurements are run at room temperature within 48 hours
of collection. pH measurements are made on an Orion 3 Star pH meter
with a glass Ross pH combination electrode. An additional pH
measurement is run in parallel, during ANC titrations run on a Metrohm
888 Titrando titration system. ANC's are run using Gran titration
methods with a titrant concentration of .02N HCl. Specific
Conductivity measurements are run on a YSI model 32 Conductance meter.
Any remaining volume after lab analyses are completed is saved and
poured off into acid washed 250 mL LDPE bottles, which are labeled and
stored in the archive building.
Precipitation samples are collected weekly at rain gage sites STA/22,
RG-1, RG-4, and RG-23. In the event that precip is falling on a sample
day, collections are postponed until the precipitation ends. A funnel/
bottle rain collector setup is used during the summer months, and 5
gallon HDPE snow buckets serve as collectors in the winter. All precip
collection gear is acid washed annually prior to field deployment.
Collectors are replaced weekly with a clean setup which has been tap
water scrubbed and rinsed in copious amounts of DIW. All precip
bottles/snow buckets are tightly capped and bagged at collection time
to prevent contamination in transit.
Since June 1, 2013 filtered, frozen samples are thawed and then
analyzed for multiple analytes (full suite listed in Table 1). All
methods used prior to 2013 are described in Buso et al. 2000 and have
been recently updated the following publication by the USDA Forest
Service, Northern Research Station 2019 (USFS Durham Lab water
analysis method detection limit (MDL) limit of quantification (LOQ),
2010 - present. Environmental Data Initiative.
https://doi.org/10.6073/pasta/890b1fadb61d3e86dc6f3d9afea79705
Throughout the history of HBWatER there have been changes in
analytical protocols as new instrumentation or new techniques became
available. The history of the record from inception to 2000 is well
documented in Buso et al. 2000 and USFS Report 2019.
Project personnel:
Current Contributors: Tammy Wooster is responsible for weekly sample
collection and sample analysis and processing in the Pierce lab. Geoff
Wilson regularly assists in the weekly rounds. Brenda Minicucci
assists with data entry, data QA/QC and routine data analysis. Jeff
Merriam is responsible for chemical analyses of samples. Emily
Bernhardt and Emma Rosi are responsible for recruiting funding to
sustain the record, approving data for publication, and data analysis
for scientific discovery. Marie Carolina Simao is responsible for the
creation and maintenance of a data visualization portal that allows
users to easily interact with the HB WatER dataset
(www.hbef.streampulse.org)
Past Contributors: For more than thirty years of this fifty year
research effort Don C. Buso led the effort to collect, filter, and
analyze samples and was responsible for data QA/QC and analysis. This
record would not exist without Don’s careful attention to detail. Many
technical support staff contributed to the collection, analysis,
management and maintenance of these long-term biogeochemical data
including Thomas Butler (1978-present), John Eaton (1965-1988), Donald
Buso (1975-2015), Phyllis Likens (1973-2014), Lorie Burkes, Jonathan
Crowdes, Larry Derosia, Brian Dresser, Yalinka Falusi, Marilyn Fox,
Steffan Freeman, Michael Hallet, Roy Hamme, Herb Karsten, Alexia
Kelley, Sue Koppes, Michael Koterba, Sherry Leis, James Lowe, Chris
Masson, Ezra Morency, Olive McGregor, Alex McPhail, Scott Nolan,
Barbara Polan, Ginny Scarlet, Denise Schmidt, Phyllis Sprague, Kathie
Weathers and Roseanne Withington. Co- Founders, F. H. Bormann, N. M.
Johnson and R. S. Pierce were instrumental in the initiation and
conduct of this overall, long-term project (The Hubbard Brook
Ecosystem Study).
Data verification: Quality analysis / quality control has been
performed as described in Buso et al. (2000). All collected data are
included in the dataset, including data that failed QA/QC controls;
data that are questionably reliable are marked as “FALSE” in the QAQC
column of the dataset, and should be treated with extreme caution.
Field Notes and Analysis Records: Field notes and field data
collection notes are transcribed into weekly data sheets. When entered
into HBWatER database, we carefully check that date, time of sample
collection, temperature, gage height, precipitation amounts, codes and
site observations are correct and within reported range. All chemical
analyses conducted at the Pierce lab (pH, ANC, Specific Conductance
and DIC) are entered into HBWatER and compared to historic and recent
data ranges of lab analyses.
All lab analyses are compiled, periodically forwarded to the the
Forest Service and archived. When chemical analyses are received from
the Forest Service lab, all data is examined for obvious outliers. A
team of the Forest Serve lab technician/lab manager, field
technician/lab manager and data entry/ manager meet and review all
results of chemical analyses. A list of reruns is complied of possible
contaminated samples or errors in analysis. When reruns are complete
and database is complete the new data are reviewed by the lead
investigators prior to posting. All field notes are archived at the
Hubbard Brook Forest Service office.
Citations
Buso, D. C., G. E. Likens and J. S. Eaton. 2000. Chemistry of
precipitation, stream water, and lake water from the Hubbard Brook
Ecosystem Study: A record of sampling protocols and analytical
procedures. General Technical Report NE-275. Newtown Square, PA., USDA
Forest Service, Northeastern Research Station. 52 pp.
Merriam, J., W. H. McDowell, and W. S. Currie. 1996. A
High-Temperature Catalytic Oxidation Technique for Determining Total
Dissolved Nitrogen. Soil Sci. Soc. Am. J. 60:1050-1055
