Details of the
Hubbard Brook landscape Biomass programThis page contains Tom's notes about the details of the biomass programming for Watershed 6. They are here as much for our remembrance of what we did as for the benefit of anyone else who needs to know the minutia.
for biomass:
zones for eqns are old-style elevation zones. each plot uses eqn appropriate to elevation, but zone summaries are by vegetation zone.
explain which species use other species productivity eqns and chemistry
different years use different sets height eqns
productivity calculated by estimating diameter of tree the year before (from annual ring incr eqns--curve derived from avg of last 5 years??) and doing 2 estimates of biomass from given year and previous year. subtract last year's mass from this years mass to get productivity. for leaves and twigs, the two masses are averaged. older needles not included. dead branches are not included in the calculations.
dead tree decomp factors (Echmiers)
species unkn and black cherry are calculated with sugar maple eqns
chem is same at all elevations for the moment
biomass for each tree in txt file: and species >14 uses SM eqns; total abvgrd uses separate eqn for total, not just sum of parts as in interactive (is this true and why there is discrepancy? no i think it sums the same way in both, not sure why different)
code 2--we assumed they were there all along. get a zero in year discovered and 2 for years before it was discovered.
for bird--used mid ht eqns and a combined m+low pv eqns (whittaker unpublished).
older needles--formerly we changed whittaker coeffs for current leaves and twigs to include older needles (a=0.6509; b=0.6137). for this program we reverted to coeffs in paper for current year needles only and added loop to calculate older needles separately using coeffs from unpublished whittaker eqns.
need to better document eqn files in header and explain codes in an eqn page. make files available as data.
for bird area--use average of mid and low estimated height; use special mid+low set of allometric eqns for plant parts
There are a lot of quick notes:
1. Snags are "standing dead trees" without branches but with a bole left above dbh. Dead trees are "standing dead trees" which still have most of their branches.
2. "All live trees" includes healthy and sick trees. "All dead trees" includes standing dead and snags.
3. Only live trees were done in 1965 with no special notation for sick trees.
4. Not all size classes were done in all years, but the >10 cm live tree class was done in all years.
5. In 1965, trees >1.5 cm dbh were measured on a 10 x 10 m plot nested randomly within each of the 25 x 25 m grid units.
6. After 1965 all trees >10 cm dbh were measured on each of the 208 25 x 25 m grid units - all the trees on the watershed were measured.
7. In 1982 and 1987, trees 2.0 - 9.9 cm dbh were measured on a stratified random set of 25 x 25 m grid units; on 32 units in 1982 and on 35 units in 1987.
8. In 1992, 1997 and 2002 the 2.0 - 9.9 cm trees were measured in a 3 x 25 m subplot in each of the 25 x 25 m grid units.
9. You should study the year-specific data documentation for the special details and access to the original raw tree by tree data.
10. There are several ways of selecting plots which are redundant, and all the various selection processes are here because the author was trying to learn how to create all these options. The answers should cross check - so if you select plot 44 in option 2 and or enter plot 44 under the single plot option you should get the same results!!
More detailed notes:1. There is some "fudging" in the calculations involved with filling out the 2 cm and 10 cm size classes for every survey year except 1965 (see Footnote 6). In 1965, the field dbh measurements were recorded to the whole number in cm. We presume that this really means that a 1.5 cm tree would have been recorded as 2 cm and a 2.4 cm tree also as 2 cm. However, in the following years all measures were recorded to the tenths decimal (e.g. 2.4, 31.7 etc.), and in the >10 cm tree survey, only the trees >10.0 cm dbh were measured--trees 9.5 to 9.9 cm were not measured. Similarly when we were measuring the 2 to 9 category after 1965, no trees 1.5 to 1.9 cm were measured. Thus, in making calculations with data after 1965, the 2 and 10 cm size classes will come up short because half of each of these size classes was not included. To avoid this we have taken the liberty to assume there are just about as many trees in the 1.5 to 1.9 range as in the 2.0 to 2.4 range and every time a tree occurs in the 2.0 to 2.4 range it is doubled. The same is true for trees 9.5 to 9.9 and 10.0 to 10.4. This is trivial in the way of things but may crop up in stem counts not quite coming out right in some combinations of requested analyses, and care must be taken in selecting a diameter range that splits the 9.5 to 10.4 range or the 2.0 to 2.4 range (again, see Footnote 6). From 1992 onwards, there is sufficient data to see how close this assumption is since all plots had 2.0 - 9.9 cm dbh trees measured in a subplot. Using the data file, one could count the number of stems for a species in the 10.0 - 10.4 range and then the number in the 9.5 - 9.9 range (correcting for sample area differences) and see how incorrect it would have been to make the doubling assumption noted above.
2. In designing the diameter distribution per hectare graph, again some "fudging" was necessary to even out diameter classes. The program lumps all the trees that belong in each 1 cm diameter class (e.g. 3.5 to 4.4 for the 4 cm class, 4.5 to 5.4 for the 5 cm class, etc.), and when run with a standard size class choice, the lower ends of the range (either 2 cm or 10 cm) were already doubled to account for missing trees as explained above (and in Footnote 6). However, if the "select a diameter" option is used and the selected lower or upper diameter is an integer, then really only half of the diameter class for these endpoints is included in the calculation (e.g. 20 to 30 includes only 20.0 to 20.4 for the 20 cm class and 29.5 to 30.0 for the 30 cm class). So that the 1 cm classes on the ends of the selected range don't come up short, the program doubles the stem count for the lower diameter and upper diameter values (unless they are 2 or 10, because these were already doubled earlier in the program). This doubling is based on the assumption that there will be about the same number of stems in either half of a 1 cm diameter class. It was assumed most users will enter integers in the "select a diameter class" option, so this only works if the entered value is an integer. If the selected range is 19.5 to 30.4, nothing will be doubled, but because this choice includes whole 1 cm diameter classes (19.5 to 20.4 for the 20 cm class and 29.5 to 30.4 for the 30 cm class), the graph will be correct anyway. However, if the selected range is 19.8 to 30.1, nothing will be doubled and both the 20 cm and 30 cm diameter classes will look short in the graph (19.5 to 19.7 trees are not accounted for, nor are 30.2 to 30.4 trees). Therefore, it is recommended to use either integers (except 10; see Footnote 6) or whole 1 cm diameter classes (X.5 to X.4) in the "select a diameter" option.
3. Remember that in 1965 a plot was a 10 x 10 m plot within each 25 x 25 m grid unit. From 1977 on, all plots for trees >10 cm were 25 x 25 m plots (the entire grid unit). Thus one cannot expect to compare plots between, say, 1977 and 1965. In addition, although considerable effort was expended, trees on the plot lines may "drift" from one plot to the next between years because the observer estimated the center of the tree to be on one plot in one year and in the adjacent plot the next time (see Footnote 5). So there are situations where the basal area of a plot may jump or drop considerably because a single large tree is right on the line and may have ended up in one plot in one year and another plot the next time. Again, this is trivial when summarizing over many plots. I suppose occasionally a tree on a watershed boundary plot may have suffered a similar fate, so one year the watershed has quite a bit more biomass because a single large tree was out one year and in the next. All this is now solved because the trees were all tagged and individually identified in 2002.
4. In 1982 and 1987, trees 2 to 9 cm dbh were not measured on all plots; only on 32 of the 208 25 x 25 m plots in 1982 and on 35 plots in 1987 (although the 2 to 9 cm trees were measured on the entire 25 x 25 m grid unit in these years, rather than just on a subplot as in subsequent years). When running 1982 and 1987 for trees 2 to 9 cm, the list of plots will show up as the entire list or group you have selected but in reality the program will only include those plots within this group which have data for the 2 to 9 cm trees. It is also possible to select a group of plots you want to do for which none had the 2 to 9 cm trees measured - you will get no results in this case. (Click here for a map of the small tree plots for these years.)
5. The option for "zones" corresponds to three general areas of the watershed which are loosely defined forest community types. The "spruce-fir" is on the ridge and is largely dominated by red spruce, fir and white birch - it comprises 26 of the 25 x 25 m grid units (zone 1 in the raw data file). The "high hardwoods" includes 55 plots which are on the ridge and shoulder of the ridge and is mostly beech with some sugar maple and birch (zones 2 and 3). The "lower hardwoods" are sugar maple, beech and yellow birch and include 127 plots (zones 4 and 5). This latter area is probably more typical of the "Northern Hardwood Forest" than the "high hardwoods".
6. The following table includes tree species found in our vegetaion surveys.
# Acronym Common name Scientific name 1 ACSA Sugar maple Acer saccharum 2 American beech Fagus grandifolia 3 Yellow birch Betula alleghaniensis 4 White ash Fraxinus americana 5 Mountain maple Acer spicatum 6 Striped maple or moose wood Acer pensylvanicum 7 Pin or fire cherry Prunus pensylvanica 8 Choke cherry Prunus virginiana 9 Balsam fir Abies balsamea 10 Red spruce Picea rubens 11 White or paper birch Betula papyrifera 12 Mountain ash Sorbus americana 13 Red maple Acer rubrum 14 Eastern hemlock Tsuga canadensis 15 unknown, used for unidentifiable rotten snags 16 Quaking aspen Populus tremuloides 17 Black cherry Prunus serotina 18 Shadbush Amelanchier sp. 19 Big-tooth aspen Populus grandidentata 20 Willow Salix sp. 21 Alternate-leaved dogwood Cornus alternifolia 22 Cherry (unspecified) Prunus sp.
For any further questions, please email thomas.siccama@yale.edu.
Web page created May 2004
by Thomas Siccama and Ellen Denny