FED/MAC Data

Leaf Area Indices

During the summer of 1990, estimates of leaf area indices [LAI] were gathered at several sites within the NASA Goddard Code 923 Forest Ecosystems Dynamics (FED) research area in the International Paper Experimental Forest west of Howland in central Maine. The following document details the locations, methodology, and results of those measurements.

Methods - Instrumentation

The LI-COR LAI-2000 leaf area index meter [LI-COR, Lincoln, NB] was used to estimate LAI in the FED research areas. This device uses two sensors, one reference ('above' or 'A') sensor set up in an area with unrestricted view of the sky, and a second, ('below canopy' or 'B') sensor that is moved through the vegetation to acquire within canopy measurements. The reference sensor was programmed to acquire horizon-to-horizon incident light levels at 30 or 60 second intervals during the time that the within canopy sensor was in use measuring light levels at known locations within the forest. The data for both sensors is recorded with a time stamp on attached data loggers and then dumped to a computer through an integral RS-232 port for analysis. The data logging units may also be joined with a cable and the data combined and analyzed in the field, but this option was not used this season.

Field Use

The 'A' or 'above' canopy unit was installed in a large clearing as near as possible to the area to be sampled for LAI. The nominal sites had no vegetation above 15 degree from the horizon, the viewing window of the sensor. Ideal conditions would include uniform overcast skies, but if there were bright sky areas or any direct sun, a 'mask' was installed on both sensors to occlude the 90 deg. segment of the sky which included the sun. When the mask was in place the 'within' or 'B' canopy sensor had to be used in the same orientation as the reference sensor to reduce errors due to differential response of the exposed areas of the detectors.

The 'B' canopy sensor was used in one of two ways, depending on whether or not the units were masked against bright sky conditions. When not masked, four replicate samples of within-canopy light level were made at each LAI position by pointing the sensor wand to the north, east, south and west at shoulder level, with a datum being taken at each of the cardinal directions. When masked, the wand was always held in the same orientation (compass bearing - usually pointed magnetic north) within the vegetation and three replicates taken across the LAI position in a north-to- south direction over a distance of approximately 3 meters. Two LAI measurements were taken at some locations where a heavy understory vegetation was present. These were taken above the understory and at the forest floor. Such data will allow the differentiation of 'canopy' and total LAI (including understory) for those sites.

Data Analysis

The software provided by LI-COR was used to download the data loggers (COMM package) and to reduce and analyze the data (C2000 package). An intermediate step was taken to check the user calibration of the sensors. The COMP X program of the data loggers allows the cross-calibration of the sensors. In addition, after the COMP X calibration, 5 to 10 additional measurements were made with the sensor heads both seeing the same 'sky' so that a linear regression could be fit between the 5 sky band counts of the 'A' and 'B' sensors. Using the slope of the regression the 'A' sensor counts were multiplied so that they corresponded as closely as possible to the response of the 'B' instrument to like light levels. In most cases this calibration amounted to only a few percent difference from the COMP X calibration. However, when using the masks differences of up to 30% were found. This is probably due to a difference in size of the hand-cut 90 degree masks and the marked difference this can make on the relatively small area of the nadir sensing element.

Once the 'A' sensor response was corrected for possible calibration errors the C2000 software was used to combine the 'A' and 'B' sensor readings, compute their LAI, mean tip angle (MTA) and total diffuse sky (DFN) values and output the data to a spreadsheet program for further analysis, tabular and graphic output.

Some users of the LAI-2000 (personal communication, S.T. Gower) have reported under estimation of LAI of up to 30% in heavily clumped canopies (especially conifers). In this initial data report no correction is made in the Howland data for potential LAI errors. The potential error involved due to canopy element clumping will be analyzed in the near future.

Sites Measured - Summer 1990

Measurements were made in early July and early September around the Howland research site. This time period encompasses the peak season when deciduous trees were fully leafed out and prior to any fall leaf loss.

Hemlock Site [12 July] - At the hemlock site the new line was surveyed with the LAI meter, including the 125N, 175W point, sites 46, 44, and 42A. The line was also measured at 10 m increments from 42A back to the road to increase the resolution of the sample and check the hysterisis of the measurements. Four replicates were made at each site. Several sites were resampled (originally done in 1989). These included 0N, 0W through 0N, 175W and 25N, 0W through 25N, 175W.
Sruce Site [15 July] - LAI estimates were made at the sites of new prisms in the spruce site, 25W X 25N, 25W X 125N, 175W X 25N, 25W X 175N, and 125W X 125N. Others were resampled as well. Esker Site [16 July] - A new series of lines was run at the north end of the esker, adjacent to the southern boundary of the gravel pit. These sites coincide with six new prisms in the predominately deciduous site. The prism sites are 25 m apart, run on a due (magnetic) south line. LAI's were taken at each of the prism sites, between the sites (at ca. 12.5 m intervals) and in a northern line parallel and 25 m to the west of the original line, also at 12.5 m intervals. Four replicates were made at each site.
Cal Site [16 July] - A single line was run through a regenerating hardwood 'island' next to the Cal site. Trees in this area are small, to approximately 3 meters. The samples were made approximately 12 m apart on a 200 deg. line. Four replicates were made at each of 15 sites.
Townline [05-12 September] - The Howland/Edinburg townline was sampled to the east and west from Gunbarrel Road at 15 meter intervals. LAI measurements were made at the 15 m points and 10 m to the north and south of the line. Four replicates were made at each point, except for a portion of the line to the west where the mask was used and three replicates made per point. The data file has been numbered with x coordinates (east-west transect line) from 0 to 1012 meters and y coordinates of 0, 10 or 20, referring to on the townline (10m), 10 m south (0 m), or 10 m north (20m). Whenever understory formed a significant portion of the LAI, a canopy (shoulder high and above) and a total (at bottom of canopy, on forest floor) LAI were both measured for a point. Theseare identified as C and T in the table.
Alders [13 September] - Three LAI points sampled in the alder stand to the east of the bridge on the esker road.
Parabola Site [12-13 September] - A grid was sampled for LAI with the parabola tower as its center point. The grid is 100 X 100 m, with LAI estimates made at 10 m intervals. The grid runs roughly north/south, with its origin in the northwest corner of the grid.
Parabola Tower [12 September] - A vertical transect of LAI was made up the parabola tower at 2 to 5 m intervals. Each sample is identified by its location in meters from the bottom of the canopy.
Larch Plantation [13 September] - Two parallel north/south LAI transects were run 20 m apart in a larch plantation on Gunbarrel Road. LAI samples were made at 10 m intervals along the two lines. Three replicate measurements were made at each point.

Data Available:

Contact:
William T. Lawrence
Department of Geography
University of Maryland
College Park, MD 20724
(301) 405 6809
blawrence@glue.umd.edu