Hubbard Brook Ecosystem Study Hubbard Brook Ecosystem Study ...From here on you are going to have to actually read stuff in order to learn anything, sorry about that!
I say "how we have" and not "how to" calculate etc. This is because I would guess that the next person to take over will change things- all well and good so long as you can get the same answers from the same data before improving things. I would guess that the next person will go to new things like EXCEL or such spread sheets rather than BASIC and raw ASCII files of the data. Using mice to calculate moles.
1. The forest population inventory
(See the figure)
Figure 1 for a map of WS 6.
The methods of sampling the forest has varied slightly over the several inventories.
The plots on WS-6 are not reproducible sampling units. That is you cannot calculate and compare the biomass of plot 14, for example, between sampling times. That might seem like quite an absurd statement for Hubbard Brook science, but it is true. The pro blem lies in plot edge trees. The trees are not individually tagged or marked so trees on the edges of plots can "drift" between plots between years. Our rule is the center of the tree has to be "in" and this is a wishy-washy call when sighting along b etween 25 m grid stakes at 4 PM on a hot buggy July afternoon when racing a thunderstorm. Also, for the period up to 1987, the corners had some stake problems. So if a 40 cm dbh sugar maple "left" a plot in 1982 and "returned" in 1992 the plot would look like it gained quite a bit. Since the plots are all contiguous (except edge plots) and the summaries are inclusive of large numbers of plots (thirds or subareas) this drift is not a problem for area totals. WE do make a special effort to deal with trees on the edge of the watershed (edge plots), but even then I know we lose and gain an occasional tree. We did make an effort to deal with this in1987 by trying to making maps of each 25 x 25 m plot and noting the location of all potential "problem trees". (We finished 3/4 of these maps and they included a lot of things such as boulder fields, streams, gaps, fern glades etc. The maps still exist and have been completed in 1992 for the remaining 1/4 of the plots in the east side beech hell area (upper 2/3 ar ea next to WS 5). The idea was to use the detailed data to make GIS stuff which never really got done, but still could be.) We noted the species, dia location and which plot each line/edge tree was assumed to be in 1987. We tried to use this in 1992, but it seemed to be too much of a pain in the ass to do for several reasons which I don't remember now. So, we plodded onward in our old ways.
In 1965
The watershed was surveyed into 208 25x25 m permanent grid units. Corners were marked with aluminum welding rod- a very bendable material. Within each grid unit a 10 x 10 m plot was randomly located, but not permantly marked. Diameters of all trees > 2
cm dbh were measured and recorded to the nearest cm. Smaller diameter class stems were sampled in sub plots located in the 4 corners of the 10 x 10 m plot. Stems taller than 50 cm and less than 2 cm dbh were counted on 4 2 x 2 m plots, one in each corne
r of the 10 x 10 m plot. Stems < 50 cm tall were counted in 4 1 x 1 m plots located in the corners of the 10 x 10 m plot and within the 2 x 2 m plot. Dead stems were not measured in 1965. This sampling intensity resulted in a 16% area wise sampling of
WS-6. (Super note: It is interesting to note that we actually forgot a subset of the biomass in our original biomass paper. That is the biomass of the trees less than 2 cm dbh. The "current growth" of these were measured and are included in the "Herb" pa
per, but the total biomass of this lessor subset got overlooked. I doubt that it amounts to much but no one ever caught the oversight).
In 1977
In 1977 the forest inventory was expanded to a 100% inventory of all trees live and standing dead > 10 cm dbh. The sampling units were the 208 25 x 25 m grid units(plots). Dbh's were recorded to the nearest mm- although fully aware that this is not real
ly an increase in accuracy given the fluff in dbhing trees. Dead stems were noted in two types based on the amount of above ground stem and branches left on the dead tree. A "dead" tree was sufficiently and recently dead such that most or a lot of crown
branches were still on it. A "stub" was a dead tree which was broken off above dbh and thus was just a bole remnant. Dead trees leaning at more than 45o were not included in this measure. See EDEX______.
We also sampled the seedlings and saplings in 1977 on the lower 24 grid units. This came about because in 1977 the original intent was to resample the watershed as was done in 1965. It became obvious immediately (after one days work) that it would be far
more usefull to do an entire inventory. (One reason such an inventory was not done in 1965 was that computers were sort of a dream on the horizon and too much data could be problem.) So we measured the lower 24 grid units using the nested 10 x 10 m plot
system. This subset of measured on these 24 plots has been contuued in all subsequent samplings.
In 1982
In 1982 the inventory of stems > 10 cm dbh, live and dead, was done in the same manner as was done in 1977.
In 1982 we expanded the inventory to include the 2-9 cm diameter classes over the whole of the watershed by sampling this size class in total on 32 of the 25 x 25 m plots. These plots were chosen at random such that the whole of the watershed would be r
epresented. We did this by using one random plot on each of the horizontal rows of plots (see map in EDEX ____). Rows with one or two plots were not included and long rows may have 2 plots. Live and dead trees were done for this size class as well as t
he larger trees.
In 1982 the lower 24 grid units (plots) were sampled for seedlings and saplings as described for 1977
In 1987
In 1987 the inventory of stems > 10 cm dbh live and dead was done in the same manner as was done in 1977 and 1982.
In 1987 the 2-9 cm dia classes were sampled in the same manner as was done in 1982 except that 2 more plots were added such as to better represent the spruce/fir area on the upper east side of WS-6. This area was not adequately sampled for the 2-9 cm di
a classes in 1982, because the "random" plots for the upper rows did not happen to fall in the spruce/fir area. See EDEX_______.
Seedlings and samplings on the lower 24 plots were not done in 1987 rather they were done two years later in 1989. We did not have the crew/time in 1987 to finish this aspect of the sampling.
In 1992
In 1992 the watershed total inventory of trees live and dead > 10 cm dbh was done as described for preceding samplings.
In 1992 the 2-9 cm dia classes were sampled with a subplot on each of the 25 x 25 m grid units rather than by a total inventory or a random subset of these units. Each subplot consisted of a 3 m wide strip along the base of each 25 x 25 m grid unit. A 3 0 m tape was stretched tight between the corner rebars and then a 3 m piece of rope was used to check each tree which was near the 3 m distance up from the tape to see if it was "in". In a few cases where the lower edge of the plot would put the subplot i n a road, trail or stream, one of the sides of the plot was sampled. It was noted on the original field sheets which side was used. See EDEX_____
The length of the side was also recorded because it is often > or < 25 m due to "errors" in the layout of the grid or stakes which have "drifted" over the years. Stakes fell over/out and were put back w/o remeasuring.
The seedling and sampling were again sampled on the lower 24 plots in 1992.
From this extensive list of equations and data we have figured out and extracted for use the key ones for estimating biomass by species, third of the watershed and by plant part. That is, we have found the equations which produced the biomass tables in the 1974 Whittaker et al. biomass paper. These original data and equations are critical to being able to recreate the biomass estimates carried out for the 1965 forest. This was especially critical for us since it was necessary for us to recalculate the original biomass by size classes different from those used by Whittaker in the paper. Here I emphasize that the published equations are not the equations used in making the published estimates. This is explained as follows and has nothing to do with the "errors" we have found and corrected(Siccama et al. 1994).
Whittaker did dimension analysis on 7 trees of each species on each third of WS-6. Allometric equations for all the tree dimensional parameters were prepared for each species for each third of the watershed. These data and equations are in the Phase 9 printout (the reduction xerox paper work disussed above). Whittaker used these species specific and third specific equations together with third specific height data (explained later) and the third specific plot inventory data to calculate the biomass of WS 6.
So, what are the published equations? and why are they different from the ones he used to calculate the biomass?? Whittaker combined the data for the individual species for the lower 2/3. He knew that the allometry for the trees from the lower and mi ddle thirds were not significatly different. Thus the published equations for a particular parameter is based on 14 trees of each species rather that just 7. You might say this should not make any difference and you are right in the "big picture"- but it does make a measurable difference in our ability to recreate what Whittaker actually used to estimate the published biomass in 1965. For the upper third Whittaker published equations which combined the 3 hardwood species into what he called "upper har dwoods".
So given we now have the right equations, which are the parabolic volume based equations, what happens next? Whittaker knew that he had selected his trees for dimension analysis- they were not random trees. In order to link this to the unbiased forest inventory, Whittaker had John Eaton (+ some trukey I no nothing about) go out and determine the tree heights (and diameters) of 40 to 60 individuals of each species in each third of the watershed. In order to calculate parabol;ic volume of a tree, an e stimate of its height is required. Thus for the 1965 biomass estimates the height of each tree on each plot was estimated using the appropriate height equation (by species and third). From this height and diameter the parobolic volume was calculated and t his was used in the parabolic volume based allometric equations to estimate the plant biomass by part (leaves, bark, wood etc.). Also note that all these equations are in the log/log transform (base 10). Also note that Whittaker did not publish an equatio n for the biomass of the older needles of the spruce. The published equation is only for current needles and twigs. However the "older needles" equation is in the Phase 9 data and we have found it and use it.
There are other aspects of my BASIC program which are also important such as what did we do for species for which no equations were developed? Also for fir, which was not analyzed by Whittaker, we did our own dimension analysis for this species at the top of WS-5 in 1984 for bole wood and bole bark. This was a bit of opportunistic science since there was cutiting going on anyhow. This study of fir was designed to see how well the assumption that spruce equations are aapplicable to fir - an assumption made by Whittaker in the original study. Also, fir is a much more abundant species on WS 6 than is spruce. The equations for fir bole wood and bark prepared by us are now used in the computer program for these parameters for this species. Roots, leave s, branches of fir are still estimated using the spruce equations.
Table 1 accompanying this package is the listing of all the equations used in the current program to estimate biomass at Hubbard Brook. The notes and page numbers on the right are the page numbers in Whittakers Phase 9 print out on which you will find t he original equations.
Table 2 accompanying this package is a listing of the 1965 Eaton tree height equations for each species for each third. Notes on the right indicate equations used to estimates heights of species for which no data had been obtained.
Some other notes and comments on the BASIC program which calculate trees in the 10 cm diameter classes the biomass.
