NPP Study: Reference harvest data

Study number: 


Data set ID: 


Data Access: 


Dataset status: 


This is the reference harvest biomass data of plants near, but outside the grid of permanent NPP quadrats that was harvested for each of 15 sites. Height and cover are recorded in the field. Live biomass is weighed in the lab and all measurements are recorded as reference harvest data. The NPP sites are grids of permanent 1 square meter quadrats established in 15 sites: three sites in each of 5 community zones (grama grassland, creosotebush scrub, tarbush flats, mesquite dunes and playa). Grids consist of 49 quadrats arranged in a square 7 x 7 pattern, with quadrats 10 m apart (P-COLL has 48 quadrats in a 3 x 16 pattern).

Data sources: 



field data sheets

Plant samples are harvested adjacent to permanent grids as described in Research Project documentation, avoiding plants obviously damaged by livestock or by previous harvests, or otherwise unrepresentative of plants within permanent quadrats. Dead biomass (other than that obviously produced in current growth season) is removed and discarded; living material is dried and weighed to nearest 0.01 g. Data are entered from the data sheets into a PC data file, using a Fortran data entry program. The data file created is named "NPPRyys.DAT" where yy = year of sample (e.g., 89 for 1989) and s = season of sample (W for winter, S for spring, F for fall). This file is then uploaded to the IBM mainframe and its name maintained. These data files are checked by a SAS program, HARVTEST (listing attached), which checks for duplicate observation numbers and for observations with missing cover, height, or weight values. The harvest data files are usually stored temporarily as SAS data sets, which are the basis for further analysis, but which are not maintained as permanent files.

Quality Assurance: 

Analysis of the harvest data proceeds by building linear regressions of plant biomass versus volume. SAS programs are used to calculate these regressions. In the first few sample dates, field observations were not consistent, and volume had to be calculated from height and diameter measurements (rather than height and cover) for some species. DWCOV SAS (sample listing attached) was modified and run for each particular species (inserting the appropriate 4-letter acronym). [When diameter, rather than cover, was measured for a species, the program, DWDIACOV SAS calculated those regressions.] More recently, DWREGALL SAS (listing attached) was used to calculate the regressions for all species in a single step, rather than having to rerun the program for each species individually. These programs read in the SAS data set created above, sort the data by zone and site, and calculate volumes for each occurrence of the species (volume = cover*height*100). They test for significant differences in the regression among sites (significance given by F-test of site*vol interaction term); they then calculate a regression of biomass versus volume for all occurrences pooled, and then separate regressions for each site. After each run of the SAS programs, one must ask for a printed copy of the listing; there is no output file or other permanent record. The linear regression is restricted to force the intercept through zero; in the rare case where such a regression will not work (occasional negative r squared values), the SAS program can be modified to remove that restriction. Hard copies of these regression outputs are saved; for each species, the best regression(s) are selected (either pooled or by site, either intercept = 0 or not) and the slope parameter noted for use in analyzing quadrat data. These parameters are entered into the regression program used to analyze the quadrat data (see documentation for quadrat data).

Additional information: 

5 vegetation communities with 3 sites in each zone: creosotebush scrub: CALI, GRAV, SAND tarbush flats: EAST, TAYL, WEST grassland: BASN, IBPE, SUMM mesquite dunes: NORT, RABB, WELL playa: COLL, SMAL, TOBO See attached site map.


Winter, spring, and fall of each year

Date Data Last Updated: 



Long-term dataset: 

Signature dataset: 

LTER Core Area(s): 

Data category: 

Research Area: 

Research Network: