This data set contains two NPP data files and one climate data file (ASCI .txt format). The NPP files contain data for above-ground biomass, litterfall, and nutrient content of above-ground vegetation, organic surface layer, and soils measured during an 18-month period in 1970 and 1971 at two contrasting tropical seasonal evergreen forests in Magdalena Valley, Colombia. The climate record provides mean monthly and annual precipitation (1951-1992) and mean monthly and annual average temperature (1970-1997) from Barranca Bermeja (7.00 N 73.80 W) near the Magdalena Valley sites.
One forest stand sits atop a perched water table on a typical valley terrace. It is low in height, basal area (22 m2/ha), and above-ground biomass (18,109 g/m2), but rich in palms with a simple two-layered structure, frequent windthrows and mortality, and few older trees. The contrasting forest stand developed on a lower slope site under more advantageous soil and water conditions. Although number of stems/ha in this forest are smaller and palms less conspicuous, the slope forest shares dominant species of both forests but with taller trees and greater basal area (32 m2/ha) and above-ground biomass (32,581 g/m2). The above-ground biomass and the vegetative bioelement stores were estimated by harvesting sample trees and palms and allometric regressions.
Above-ground net primary production (ANPP) is based on total litterfall accumulation, measured over a 12-15 month period (1,202 g/m2/yr for the terrace forest and 873 g/m2/yr for the slope forest), giving minimum estimates of NPP.
Revision Notes: Only the documentation for this data set has been modified. The data files have been checked for accuracy and are identical to those originally published in 1999.
The Net Primary Productivity (NPP) data collection contains field measurements of biomass, estimated NPP, and climate data for terrestrial grassland, tropical forest, boreal forest, and tundra sites worldwide. Data were compiled from the published literature for intensively studied and well-documented individual field sites and from a number of previously compiled multi-site, multi-biome data sets of georeferenced NPP estimates. The principal compilation effort (Olson et al., 2001) was sponsored by the NASA Terrestrial Ecology Program. For more information, please visit the NPP web site at http://daac.ornl.gov/NPP/npp_home.html.
Cite this data set as follows:
Folster, H. 2013. NPP Tropical Forest: Magdalena Valley, Colombia, 1970-1971, R1. Data set. Available on-line [http://daac.ornl.gov] from Oak Ridge National Laboratory Distributed Active Archive Center, Oak Ridge, Tennessee, U.S.A. doi:10.3334/ORNLDAAC/477
This data set was originally published as:
Folster, H. 1999. NPP Tropical Forest: Magdalena Valley, Colombia, 1970-1971. Data set. Available on-line [http://daac.ornl.gov] from Oak Ridge National Laboratory Distributed Active Archive Center, Oak Ridge, Tennessee, U.S.A.
Project: Net Primary Productivity (NPP)
Biomass, litterfall, and nutrient content of above-ground vegetation, organic surface layer, and soils were determined for two contrasting tropical seasonal evergreen forests at Magdalena Valley, Colombia, during an 18-month period in 1970 and 1971. The Magdalena Valley study site (6.39 N 73.56 W) is located 50 km south of the town of Barranca Bermeja, about 200 km north of Bogota. Annual rainfall for the region is around 3,000 mm, with maximum amounts being received in May and during October/November. The study was sponsored by the German Research Foundation.
Of primary interest were biomass and nutrient dynamics of a forest stand that had developed atop a perched water table on a typical valley terrace. Perched water tables give rise to pseudogley soils with low pH, prolonged periods of saturation, low nutrient status, reduced rooting depth, and small exploitable volume. The resulting forest is of relatively low biomass and stature, with frequent windthrow and mortality, and few older trees. Biomass and nutrient estimates for a contrasting forest stand on a lower slope site, which had developed under more advantageous soil and water conditions, are included in this data set for comparison.
The two sub-sites differed markedly in basal area (22 m2/ha terrace forest; 32 m2/ha slope forest). Total litterfall, measured over a 12-15 month period, was 1,202 g/m2/yr for the terrace forest and 873 g/m2/yr for the slope forest, giving a minimum estimate of above-ground net primary productivity (ANPP) (Folster et al., 1976; Scurlock and Olson, 2012). Estimates of above-ground biomass and nutrient content of biomass components and soils are also reported. Climate data are available for Barranca Bermeja (7.00 N 73.80 W) near the Magdalena Valley sites for the periods 1951-1992 (precipitation) and 1970-1997 (average temperature).
ANPP, BNPP, and TNPP values reported in Olson et al. (2012a, b) and Clark et al. (2001a, b) differ from values presented herein due to different calculation methods (Table 1).
Table 1. ANPP, BNPP, and TNPP values reported by various published data sources
| File Name or Description | Data Source(s) | Sub-Site | ANPP | BNPP | TNPP |
|---|---|---|---|---|---|
| gC/m2/year | |||||
| mgd1_npp.txt | Folster et al. (1976)1
|
mgd1 mgd plot_BPa (Terrace forest) |
601 | NA | NA |
| mgd2_npp.txt | mgd2 mgd plot_BPb (Slope forest) |
437 | NA | NA |
|
NPP_Multibiome_EnvReview | Scurlock and Olson (2012) based on Folster et al. (1976)1 |
mgd (average of the two sub-sites) |
519 | NA | NA |
| GPPDI_ClassA_NPP_162_R2.csv | Olson et al. (2012a);Clark et al. (2001a)2 based on Folster et al. (1976) |
Class A 44 (MI 46) (Terrace forest) |
962 | 674 | 1,636 |
Class A 44 (MI 45) (Slope forest) |
747 | 524 | 1,272 |
||
| EMDI_ClassA_NPP_81_R2.csv | Olson et al. (2012b);Clark et al. (2001a)2 based on Folster et al. (1976) |
Class A 44 (average of the two sub-sites) |
855 | 600 | 1,455 |
Table 1 in Clark et al. (2001a)
| Clark et al. (2001a)2 based on Folster et al. (1976) |
Magdalena terrace |
960 | 190-1,150 (av 655) | 1,150-2,120 (av 1,635) |
Appendix A in Clark et al. (2001a)
| #2 (Terrace forest) |
NA | NA | 1,635 |
|
#1 (Slope forest) |
NA | NA | 1,275 |
||
tropfornpp.csv | Clark et al. (2001b)3 based on Folster et al. (1976) |
Colombia - Magdalena (BS 16) |
473 | NA | NA |
Colombia - Magdalena (terrace) |
447 | NA | NA |
||
Colombia - Magdalena (slope) |
338 | NA | NA |
||
Notes: NA = Not available. MI = Measurement ID number. The differences in NPP values reported in this table are mainly due to differences in calculation methods, as explained in these notes. Please consult original references for details. Revised data sets (R1, R2, etc) are accompanied by ORNL DAAC Data Set Change Information files. Please see the corresponding documentation for reasons why the data values were revised. 1For this table, NPP data from the original data source were converted from grams of dry weight per meter square per year to grams of carbon per meter square per year using a conversion factor of 0.5. The ANPP estimate is based on field measurement of total litterfall accumulation, giving a minimum estimate of NPP. 2Clark et al. (2001a) used a different approach to calculate net primary production values. ANPP was calculated by summing reported above-ground biomass increment + reported fine litterfall + estimated losses to consumers + estimated VOC emissions. BNPP was calculated by summing 0.2 x estimated ANPP for a low BNPP estimate + 1.2 x estimated ANPP for a high BNPP estimate. TNPP was calculated as the range between the low and high values of ANPP + BNPP. Average BNPP and TNPP estimates were also calculated. See Clark et al. (2001a) for a discussion of calculation methods, including how unmeasured components of ANPP were estimated and the basis for setting bounds on BNPP. 3ANPP estimates in Clark et al. (2001b) exclude twig and branchfall reported in Folster et al. (1976).
Site: Magdalena Valley, Colombia
Site Boundaries: (All latitude and longitude given in decimal degrees)
| Site (Region) | Westernmost Longitude | Easternmost Longitude | Northernmost Latitude | Southernmost Latitude | Elevation (m) |
|---|---|---|---|---|---|
| Magdalena Valley, Colombia | -73.56 | -73.56 | 6.39 | 6.39 | ~200 |
Site Information
The study site includes two sub-plots, a forest stand that had developed atop a perched water table on a typical valley terrace; and a contrasting forest stand on a lower slope site. The forest on the terrace flat (site BPa) is an evergreen seasonal forest, low in height, biomass and basal area, and rich in palms. The plateau site of the terraces for which these stands are typical, is characterized by soils of low bioelement status, by long phases of high water saturation, enforced concentration of roots on the surface of the mineral soil and the associated accumulation of an organic layer, and irregular phases of drought. It is believed that frequent windthrows and drought result in high mortality, preserve a youthful stage of trees in the primary forest, and strengthen the competitive position of palms.
The terrace forest has a simple two-layered structure. There is a thick continuous canopy with a height between 6 and 14 m (average maximum diameter of 15 cm dbh), and a second more open story with a maximum height of 30 m (maximum diameter 1970 of 65 cm, 1971 of 50 cm), 15% (13% and 17.6%, respectively) of all trees were palms with stems. One species, Maquenque (Oenocarpus sp.) does not grow higher than 10 m., the other, Milpesos (Jessenia polycarpa), reaches 23 m. in height and, thus, overlaps with both non-palm stories. The basal area of trees with > 10 cm dbh was relatively low (22 m2/ha).
On the steeper slopes which are associated with the dissecting water courses or the strongly dissecting scarps between terraces of different levels, ecological conditions are different as lateral seepage exercises a strongly differentiating impact apparently stabilizing the water supply downslope. Height and biomass of the forest consequently increase downslope without conspicuous change in the soil mantle.
The slope forest stand (site BPb) differs markedly from the terrace forest in regard to basal area (32 m2/ha) and total biomass. The number of stems is smaller (600 trees/ha of dbh >10 cm), the trees are taller, and the palms less conspicuous. The slope forest occupies an intermediate position between the terrace forests and forests that develop in a low hilly transition zone between the terraced Magdalena Valley and the Andean foothills. The slope forest shares dominant species of both forests: Leche perra (Pseudolmedia riglda) with the former, and Sapan (Clathotropls brachipetalum, Perillo negro (Couma macrocarpa), and Carbonero (Parkia multijuga) with the latter.
Climate data are available from the weather station at Barranca Bermeja (7.00 N 73.80 W) 50 km north of the study sites at an elevation of 130 m.
The inventory plots were 50 m x 50 m for the terrace forest (BPa; 1970) and 50 m x 40 m for the terrace and slope forests (BPa & BPb; 1971). Sample trees and palms were harvested from BPa for allometric calculations of total biomass. Understory growth was harvested on 100 m2 plots. Litterfall was collected in 0.75 m2 traps, with six traps per stand. Soil samples were extracted from 6 auger holes on two plots at each stand. Soil plot size was 4 x 4 m to 10 x 10 m; samples were taken at 10 cm intervals down to a depth of 120 cm (average). Sampling of the organic surface layer was done with 20 x 20 cm frames.
Vegetation inventories were carried out on the terrace forest in 1970 and on the terrace and slope forest in 1971. Harvests of sample trees and palms and understory were carried out on the terrace forest in 1970. Litterfall was collected over 12-15 months (May through April or June) during the 1970-71 period. Soil samples were extracted in 1971 from both sites. Precipitation data are available from 1951/01/01 through 1992/12/31. Temperature data are available from 1970/01/01 through 1997/12/31.
Above-ground biomass measurements and soil sampling were carried out once at each site. Monthly average litterfall was calculated based on bi-weekly collection. All biomass and NPP estimates are based on plant dry matter accumulation, expressed as g/m2 and g/m2/year, respectively. Climate data are expressed as monthly and annual precipitation amounts (mm) and monthly and annual average temperature (C). Monthly and annual mean precipitation amounts are provided for the 1951-1970 period (with some missing annual averages), and monthly and annual mean average temperature for the 1970-1997 period (with some missing annual averages).
Table 2. Data files in this data set archive
FILE NAME |
FILE SIZE |
TEMPORAL COVERAGE | FILE CONTENTS |
mgd1_npp.txt |
3.7 KB |
1970/05/01-1971/04/30 | Biomass and litterfall data and nutrient content of vegetation and soils for terrace forest (BPa) in the Magdalena Valley, Colombia |
mgd2_npp.txt |
2.3 KB |
1970/05/01-1971/04/30 | Biomass and litterfall data and nutrient content of vegetation and soils for slope forest (BPb) in the Magdalena Valley, Colombia |
mgd_cli.txt |
12 KB
|
1951/01/01-1992/12/31 | Mean monthly and annual precipitation data from the Barranca Bermeja weather station near Magdalena Valley, Colombia |
| 1970/01/01-1997/12/31 | Mean monthly and annual average temperature data from the Barranca Bermeja weather station near Magdalena Valley, Colombia |
NPP Data. NPP estimates for the Magdalena Valley sites are provided in two files (Table 2), one for the terrace forest (BPa) and one for the slope forest (BPb). The data sets are ASCII files (.txt format). The variable values are delimited by semi-colons. The first 18 lines are metadata; data records begin on line 19. The value -999.9 is used to denote missing values. Biomass and NPP units are in g/m2 and g/m2/year (dry matter weight), respectively.
Table 3. Column headings in NPP files
COLUMN HEADINGS |
DEFINITION |
UNITS |
Site |
Site where data were gathered (code refers to site identification) | Text |
Treatmt |
Study area or forest subsystem type where measurements were made; plot BPa = Terrace forest; plot BPb = Slope forest | Text |
Year |
Year in which data were collected | Numeric |
Month |
Month in which data were collected | |
Day |
Day on which data were collected | |
parameter |
Parameters measured (see definitions in Table 3) | Text |
amount |
Data values | Numeric |
| units | Unit of measure | Text |
Table 4. Parameter definitions in NPP files
PARAMETER |
DEFINITION |
UNITS |
SOURCE1 |
mean_height |
Canopy height of Milpesos palm (Jessenia polycarpa)2 | m |
p. 300; Figure 2, p. 301 |
leaflittfall |
Annual leaf litterfall accumulation | g/m2/year |
Table 4 |
twiglittfall |
Annual twig litterfall accumulation | ||
Frtlittfall |
Annual fruit litterfall accumulation | ||
Plmlittfall |
Annual palm leaf litterfall accumulation | ||
Totlittfall |
Total annual litterfall accumulation (sum of above) | ||
trunks |
Stem biomass3 | g/m2 | Table 6 |
branches |
Biomass of non-palm trees | ||
twigs |
Biomass of twigs (shoots) of understory plants + non-palm trees | ||
leaves |
Leaf biomass (sum of leaf biomass of understory woody plants + palms + other plants plus leaf biomass of overstory stemless palms + palms3 + non-palm trees) | ||
AGbiomass |
Above-ground live biomass (trunks + branches + twigs + leaves) | ||
stdead |
Biomass of standing dead trees | ||
AGTotmatter |
Total above-ground biomass (live + dead) | ||
Orglayermatter |
Biomass of organic matter in surface layer above soil | g/m2 | Table 9 |
Soilorgmatter |
Biomass soil organic matter (composite sample to 50 cm depth) | g/m2/50cm | Table 9 |
AGTotmatter-N |
Nitrogen concentration in above-ground vegetation4 | g/m2 | Table 9 |
Orglayermatter-N |
Nitrogen concentration in organic surface layer | g/m2 | Table 9 |
Soilorgmatter-N |
Nitrogen concentration in soils (composite sample to 50 cm depth) | g/m2/50cm | Table 9 |
AGTotmatter-P |
Phosphorus concentration in above-ground vegetation4 | g/m2 | Table 9 |
Orglayermatter-P |
Phosphorus concentration in organic surface layer | g/m2 | Table 9 |
Soilorgmatter-P |
Phosphorus concentration in soils (composite sample to 50 cm depth) | g/m2/50cm | Table 9 |
trunks-N |
Nitrogen concentration in tree trunks | % | Table 7 |
branches-N |
Nitrogen concentration in tree branches | ||
twigs-N |
Nitrogen concentration in tree twigs | ||
leaves-N |
Nitrogen concentration in tree leaves | ||
trunks-P |
Phosphorus concentration in tree trunks | ||
branches-P |
Phosphorus concentration in tree branches | ||
twigs-P |
Phosphorus concentration in tree twigs | ||
leaves-P |
Phosphorus concentration in tree leaves |
Notes: Data for palm litterfall, standing dead wood, and nutrient concentration in vegetation compartments are not available for the slope forest (BPb). 1All data are from Folster et al. (1976). 2Based on height distribution of Milpesos (Jessenia polycarpa) on the inventory plot (1970). 3Sum of biomass of Palm Maquenque + Palm Milpesos. 4The figures of bioelement content of BPa were used in calculating the BPb bioelement stores.
Sample NPP Data Record <mgd1_npp.txt>
Site; Treatmt; Year; Month; Day; parameter; amount; units mgd; none; -999.9; -999.9; -999.9; mean_height; 23; m mgd; plot_BPa; 1970-71; -999.9; -999.9; leaflittfall; 651; g/m2/year |
Sample NPP Data Record <mgd2_npp.txt>
Site; Treatmt; Year; Month; Day; parameter; amount; units mgd; none; -999.9; -999.9; -999.9; mean_height; 23; m mgd; plot_BPb; 1970-71; -999.9; -999.9; leaflittfall; 664; g/m2/year |
Climate Data. The climate data set is an ASCII file (.txt format). The first 18 lines are metadata; data records begin on line 19. The variable values are delimited by semi-colons. The value -999.9 is used to denote missing values.
Sample Climate Data Record
Site;Temp;Parm; Jan; Feb; Mar; Apr; May; Jun; Jul; Aug; Sep; Oct; Nov; Dec; Year mgd ;mean;prec; 52.8; 96.9; 157.7; 274.0; 298.7; 224.6; 177.7; 251.3; 293.3; 395.9; 269.6; 99.5; 2575.4 Where, |
The accumulation of biomass, or NPP, is the net gain of carbon by photosynthesis that remains after plant respiration. While there are many fates for this carbon, this data set accounts for litterfall accumulation, giving a minimum estimate of NPP.
This study was carried out to collect and compare information on stand structure, biomass and bioelement stores of primary evergreen lowland forests that had developed under different soil water conditions.
The tropical forest biomass data for the Magdalena Valley site are provided for comparison with models and estimation of NPP. Climate data are provided for use in driving ecosystem/NPP models.
Data were examined for general consistency and clarity. The above-ground biomass of the terrace stand was low for evergreen lowland forests in the tropics, while the slope forest was average (see Figure 4 in Folster et al., 1976). Stem wood, which dominated the biomass (75%), determined possible biomass differences observed in these tropical lowland forests.
Total leaf fall did not differ greatly in the two stands. The difference increases with components of greater local variation, as twigs, and palm fronds which apparently have a longer life span and were collected only 4 times in BPa but contributed considerable weight. Therefore, total litterfall data have to be treated with more caution than leaf fall data when collection time is restricted to one year as in this study.
Because of lack of corresponding bioelement concentration data from other evergreen lowland forests at the time of this study, there was little scope for comparative evaluation. The data available from tropical lowlands in Zaire, Ghana, and Central America show the N-levels in the leaves at Magdalena Valley as comparatively high; phosphorus and calcium as comparatively low; and K and Mg appear to be at par when allowance is made for specific site influences (see Figure 5 in Folster et al., 1976).
Sources of Error
Information not available.
Litterfall. Litterfall was collected for a period of 12-15 months in the two stands (BPa and BPb) by means of 6 collectors per stand (sampling area 0.75 m2 per collector). The litter was collected bi-monthly and combined to provide monthly averages.
Above-ground Biomass. Inventories of all individuals >3.5 cm dbh were conducted on the two adjoining plots of 50 x 50 m (BPa in 1970) and 50 x 40 m (BPa & BPb in 1971). Heights were measured after felling on the first plot so that parabolic diameter : height functions could be obtained. Stemless palms were counted on the 1971 plot (450/ha) while the rest of the undergrowth was harvested on 3 plots of 100 m2 each.
The above-ground biomass was estimated by means of harvesting sample trees (43) and palms (17) in the BPa stand and applying allometric regressions following Ogawa et al. (1965). Total biomass was calculated from inventory data by means of allometric functions between D2H and component weights derived from sample trees. Non-palm trees were selected in such a manner so as to cover the full range of diameter classes, to reflect the dominance of species in the diameter classes, and to represent average trees in regard to crown, lianas, etc.
Sample trees were harvested and weighed in 4 components: leaves, twigs (< 3.5 cm diameter), branches (3.5-20 cm except central shaft), and stem (central shaft and branches > 20 cm). The trunk weights of the biggest trees were calculated from the specific weight of wood samples (2 trunk disc / tree) and the trunk volume. See Folster et al. (1976) for regression equations. Below-ground biomass was not measured.
Palm biomass was estimated from 17 sample trees (10 Maquenque, 7 Milpesos). The regressions were calculated separately for the two species. Components were stems, leaves, and - in case of Milpesos - the mantle of dead leaf bases that normally covered the upper part of the stem. DBH was always measured without this mantle. All regressions are linear on log-log coordinates. Leaf weight could be directly correlated to D2H.
Bioelement Stores. Nutrient concentrations in above-ground vegetation, soils, and organic surface layers were estimated from samples collected at each site. Vegetation samples were obtained from the harvested components. The soil was sampled at 10 cm intervals down to a depth of 120 cm (average). Because of the shallow rooting depth, the data reported in this data set refer to a depth of 50 cm only. Each analyzed soil sample was a mixture of respective layers from 6 auger holes on two plots at each stand (identical with two of the three harvest plots). Plot size was 4 x 4 m to 10 x 10 m. Two composite samples were obtained for each site and depth. The contents were recalculated on a volume basis (kg/ha). Sampling of the organic surface layer was done with 20 x 20 cm frames and followed the same pattern as that of the soil, i.e., one composite sample (mixed from 6 frame samples) from each of two plots was taken as representative for a vegetation site.
Climate data. Climate data are available from the weather station at Barranca Bermeja (7.00 N 73.80 W) 50 km north of the study sites at an elevation of 130 m.
This data set is available through the Oak Ridge National Laboratory (ORNL) Distributed Active Archive Center (DAAC).
Web Site: http://daac.ornl.gov
E-mail: uso@daac.ornl.gov
Telephone: +1 (865) 241-3952
Folster, H., G. de las Salas, and P. Khanna. 1976. A tropical evergreen forest site with perched water table, Magdalena valley, Colombia: biomass and bioelement inventory of primary and secondary vegetation. Oecologia Plantarum 11: 297-320.
Olson, R.J., K.R. Johnson, D.L. Zheng, and J.M.O. Scurlock. 2001. Global and Regional Ecosystem Modeling: Databases of Model Drivers and Validation Measurements. ORNL Technical Memorandum TM-2001/196. Oak Ridge National Laboratory, Oak Ridge, Tennessee, U.S.A.
Additional Sources of Information:
Clark, D. A., S. Brown, D. W. Kicklighter, J. Q. Chambers, J. R. Thomlinson, J. Ni, and E. A. Holland. 2001a. Net primary production in tropical forests: an evaluation and synthesis of existing field data. Ecological Applications, 11(2): 371-384.
Clark, D.A., S. Brown, D.W. Kicklighter, J.Q. Chambers, J.R. Thomlinson, J. Ni, and E.A. Holland. 2001b. NPP Tropical Forest: Consistent Worldwide Site Estimates, 1967-1999. Data set. Available on-line [http://daac.ornl.gov] from the Oak Ridge National Laboratory Distributed Active Archive Center, Oak Ridge, Tennessee, U.S.A. doi:10.3334/ORNLDAAC/616
Ogawa, H., K. Yoda, K. Ogino, and T. Kira. 1965. Comparative ecological studies on three main types of forest vegetation in Thailand II. Plant biomass. Nature and Life in SE-Asia, Volume IV: 50-80.
Olson, R.J., J.M.O. Scurlock, S.D. Prince, D.L. Zheng, and K.R. Johnson (eds.). 2012a. NPP Multi-Biome: Global Primary Production Data Initiative Products, R2. Data set. Available on-line [http://daac.ornl.gov] from the Oak Ridge National Laboratory Distributed Active Archive Center, Oak Ridge, Tennessee, U.S.A. doi:10.3334/ORNLDAAC/617
Olson, R.J., J.M.O. Scurlock, S.D. Prince, D.L. Zheng, and K.R. Johnson (eds.). 2012b. NPP Multi-Biome: NPP and Driver Data for Ecosystem Model-Data Intercomparison, R2. Data set. Available on-line [http://daac.ornl.gov] from the Oak Ridge National Laboratory Distributed Active Archive Center, Oak Ridge, Tennessee, U.S.A. doi:10.3334/ORNLDAAC/615
Scurlock, J.M.O., and R J. Olson. 2002. Terrestrial net primary productivity - A brief history and a new worldwide database. Environ. Rev. 10(2): 91-109. doi:10.1139/a02-002
Scurlock, J.M.O., and R.J. Olson. 2012. NPP Multi-Biome: Grassland, Boreal Forest, and Tropical Forest Sites, 1939-1996, R1. Data set. Available on-line [http://daac.ornl.gov] from Oak Ridge National Laboratory Distributed Active Archive Center, Oak Ridge, Tennessee, U.S.A. doi:10.3334/ORNLDAAC/653