Data Set Citation:
When using this data, please cite the data package:
NCEAS 10022 : Shurin, Hillebrand & Gruner: Comparing trophic structure across ecosystems , National Center for Ecological Analysis and Synthesis , and Gruner D. 2006.
Experimental manipulations of nutrients and trophic structure across ecosystems
nceas.301.7 (
General Information:
Title:Experimental manipulations of nutrients and trophic structure across ecosystems
Alternate Identifier:ELSIE: Ecological Synthesis of Interactive Experiments
Nutrient availability, herbivore grazing pressure, and the indirect effects of top predators control the standing biomass, nutrient content and species diversity of primary producers to varying degrees across global ecosystems. For this comprehensive database, we systematically gathered data from the literature that experimentally manipulated (1) nutrients through fertilization, (2) herbivore pressure through exclusion, enclosion, or removal experiments, (3) predators through similar means, or (4) a combination of these treatment combinations. Special priority was given to studies that factorially manipulated multiple nutrients (especially N & P) or manipulated both nutrients and herbivores (bottom-up and top-down), and which measured species richness, biomass or nutrient content of autotrophic responses at the community level. In addition, the Trophic Cascades database (owned by E. Borer et al.) has been folded into this framework, as well as data from several other published analyses from restricted systems (e.g., H. Hillebrand periphyton database). The goals were manifold; however, the central objective was to compare with meta-analyses the relative effects of these experimental manipulations across global organismal functional groups and ecosystem types. To this end, the database was designed to be general and flexible in order to accomodate varying research questions, methodologies, spatial and temporal scales, and abiotic and biotic covariates reported across different studies. As of this writing (14 November 2006), the database contains nearly 2000 studies across terrestrial, freshwater and marine systems, and well over 10,000 treatment responses. Data are entered into four different tables in a relational design (study-level data, treatment/responses, local biotic covariates of response and manipulated taxa, & general/global taxonomic and functional characteristics of these taxa).
  • nutrient resource
  • nitrogen
  • phosphorous
  • top-down
  • bottom-up
  • herbivore
  • consumer exclusion
  • trophic cascade
  • marine, terrestrial, freshwater ecosystem
  • experimental manipulation
  • global
  • autotroph
  • biomass
  • species diversity and richness
  • grazer
  • fertilization
Publication Date:2006-11-02

Involved Parties

Data Set Creators:
Organization:NCEAS 10022 : Shurin, Hillebrand & Gruner: Comparing trophic structure across ecosystems
Organization:National Center for Ecological Analysis and Synthesis
Individual: Daniel Gruner
Organization:University of Maryland
Department of Entomology, 4112 Plant Sciences Bldg,
College Park, Maryland 20742 USA
301-405-3957 (voice)
301-314-9290 (Fax)
Email Address:
Data Set Contacts:
Individual: Daniel Gruner
Organization:University of Maryland
Department of Entomology, 4112 Plant Sciences Bldg,
College Park, Maryland 20742 USA
301-405-3957 (voice)
301-314-9290 (Fax)
Email Address:
Associated Parties:
Individual: Eric Seabloom
Individual: Jennifer Smith
Individual: Jackie Ngai
Individual: Matt Bracken
Individual: Stan Harpole
Individual: Elsa Cleland
Individual: James Elser
Individual: Helmut Hillebrand
Individual: Elizabeth Borer
Metadata Providers:
Individual: Daniel Gruner

Data Set Characteristics

Geographic Region:
Geographic Description:global
Bounding Coordinates:
West:  -180.0000  degrees
East:  180.0000  degrees
North:  90.0000  degrees
South:  -90.0000  degrees
Time Period:

Sampling, Processing and Quality Control Methods

Step by Step Procedures
Step 1:  

Data discovery

Studies were selected by examining the abstracts of all publications returned from searches on ISI Web of Science using the following search strings (for different research questions):

•[herbivor* or graz* or consum*] and [resourc* or nutrient* or fertili*]. We also included data from studies reported in published syntheses (Milchunas and Lauenroth 1993, Bigger and Marvier 1998, Proulx and Mazumder 1998, Micheli 1999, Hillebrand 2002, Coupe and Cahill 2003, Hughes et al. 2004, Borer et al. 2005, Parker et al. 2006) and searched both the literature cited in those papers and all subsequent papers citing the analyses.

• [herbivor* or graz* or consum*] and [resourc* or nutrient* or fertili*] and [divers* or species richness or evenness]. We also included data from studies reported in published syntheses (Proulx and Mazumder 1998, Worm et al. 2002).

•[fertili* nutrient or nitrogen or phosphorus] and included data from studies reported in published syntheses (DiTommaso and Aarssen 1989, Elser et al. 1990, Tanner et al. 1998, Downing et al. 1999, Francoeur 2001).

Step 2:  

Study selection criteria

We included only studies that reported mean community-level biomass (or nutrient status or diversity) responses of autotrophs to the manipulations of resources, herbivores or both. Single species responses were eliminated unless drawn from a monodominant community in the judgment of the original authors. Some studies allowed summing up the species-specific information to calculate community level biomass and diversity. The preferred metric was standing, dry biomass by unit area, although we also accepted proxy variables that have been shown to be highly correlated with standing biomass (such as chlorophyll concentration, ash-free dry mass, carbon mass, biovolume, percent cover, and primary production). These inclusive criteria incorporated more studies into the database and increased statistical power, while the robustness of these data were later checked with diagnostics (below; Englund et al. 1999). However, counts were excluded because organisms can vary in body size by orders of magnitude between systems, and because body size and abundance are expected to be inversely related (Cohen et al. 1993, Cyr et al. 1997).

Step 3:  

Study definition and data selection

We defined a study as a temporally and spatially distinct sample with internally consistent controls. Multiple studies could be reported from within one publication if, for instance, the same experimental treatments were performed in multiple streams with differing water quality. When multiple measures were reported over time, we used the last temporal sample in order to avoid phases of transient dynamics. Exceptions were made if some unusual disturbance affected some or all of the treatments or replicates, or if we decided that duration should be standardized within systems (done only for nutrient bioassays). In these cases, we used the most robust values by deferring to the working knowledge and intuition of the original authors. Seasonal studies were pooled if phenological changes necessitated the use of mean values over all samples instead of the final value to be more ecologically relevant.

Step 4:  

Study-level data

For each study, we used a unique study identifier linked to the citation of the publication. A complete list of variables that we recorded, is appended as Table 1. We obtained the following information for each study (variables are underlined):

1) We categorized the system as marine, terrestrial, freshwater and described the strata of the system by assigning aquatic studies to either pelagic or benthic and the terrestrial to either aboveground or belowground. Studies in wetlands and salt-marshes were more difficult to categorize. For these, we operationally defined studies addressing submersed or floating macrophytes or periphyton as aquatic (marine or freshwater), whereas studies on above-water rooted plants were termed terrestrial. We used a standardized set for habitat categories comprising agricultural; forest; herbaceous; tundra; shrubland; wetland; stream; lake; coastal hard bottom; coastal soft bottom; or oceanic. We also recorded a unique habitat description to enable more detailed descriptions.

2) We noted the location as the unique name of the experimental site together with latitude from -90 (S) to +90 (N), the longitude from -180 (W) to +180 (E), and the elevation (m above sea level, with negative values denoting depth).

3) We noted the ambient resource levels for the studied systems, always referring to unmanipulated resource levels (note: for light manipulation studies the low level was used). The ambient concentration of dissolved available nitrogen (N), total N-concentration, ambient concentration of dissolved available phosphorus (P), total P -concentration, total carbon (C) concentration were recorded (µmol g-1 for terrestrial and µmol l-1 for aquatic studies). Light was given as mean daily irradiance (µmol photons m-2 s-1).

4) Other abiotic variables obtained were pH, mean temperature over the duration of the experiment (°C), average annual rainfall during the experiment (mm), ambient net primary production NPP (g m-2 yr-1, categorized by total, aboveground or belowground in the variable NPP-type), ambient autotroph biomass (standing crop outside the experiment, g C m-2), categorized in the ab_unit column as maximum, min, mean or total biomass.

5) We categorized the study type as lab or field, where lab studies also comprise outdoor mesocosms or streamside channels or other contained environments, which lack direct contact to the original environment. The type of consumer manipulation includes exclosures, enclosures, removal, gradient, multiple, none. Exclosures are experiments that restrict herbivore access to plants , enclosures are experiments where stocked herbivores are kept with their forage, and removal experiments do not have physical boundaries, but remove herbivores by e.g. insecticides. A few studies used a natural gradient of herbivore presence/absence, or employed a ‘multiple’ setup, which represents a combination of the factors mentioned above or a split-plot design. All studies without grazer manipulation are termed none. The presence of a fertilization treatment was identified by a binary variable as 1 (yes) or no (0). For each study we gave the maximum food chain length as integer number, identifying the highest trophic level subject to manipulation (not measurement). The integer number of manipulated nutrient treatments and the total number of nutrients added were recorded. These variables only address unique main treatments, but not factorial interactions; thus, an experiment having +N, +P, and +N+P treatments has two main treatments and 2 nutrients added. A non-factorial study with +K+P and +K+N levels had 2 main treatments and 3 nutrients added. For studies using a fully supply of nutrients, e.g. osmocote, we arbitrarily assigned 21 as the number of nutrients.

6) The scale of the experiment was noted as the duration of the experiment (days), the area used for fertilization manipulation (fertilization area), and the area over which consumers were manipulated (consumer area) or volume of the experiment (litre).

7) If the study came from an existing database or meta-analysis, this was indicated in the meta_source column. Any special conditions or other notes were also recorded (note_study).

Step 5:  

Data for experimental treatments

Within the study, we obtained the following information on the manipulations done for each treatment level:

1) The genus or common name of the manipulated taxa and the actual food chain length in this treatment were recorded. A binary variable (presence) identifies whether this treatment was intended to contain the manipulated taxa, where 1 = yes, 0 = no. For studies without manipulated taxa (i.e., studies with fertilization treatments only), “no” is entered. The abundance of the manipulated taxa including the abundance units wasnrecorded, as well as the biomass and the biomass units. The diversity of the manipulated taxa and the diversity unit (as richness, species density, Shannon-index Simpson-index, Fisher’s alpha or other) were noted.

2) For resource manipulations, we used a suite of binary variables to describe which resources were added (n_add, p_add, other1_add, other2_add, light), with 1 = yes and 0 = no. For other 1 and other 2, we identified the type of nutrient element added (other1_typ, other2_typ). For all manipulations, we gave the rates of added nutrients in continuous variables (n_rate, p_rate, other1_rate, other2_rate, light_rate), with the units given in the variable fertilization units. We used a text string to list all the elements added (including P NO NH Fe Si B Mg Light Ca K S Water Mo CO2 complete(e.g. osmocote) micro(trace)).

Step 6:  

Data for responses

We recorded the responses of autotrophs using the following variables:

1) The genus name or other identifying category of the response taxa and the category of the response, which comprised biomass, nutrient status or diversity. The specific type of measurement could be i) for the biomass category: chlorophyll, ash-free dry mass AFDM, wet biomass, biomass (= dry), belowground biomass, aboveground biomass, carbon, biovolume, percent cover, net (total, aboveground, belowground) primary production per area (NPP, ANPP, BNPP), ii) for the diversity category: richness, evenness, index, or species density; iii) for the nutrient status category: tissue nutrients. The unit of response was also recorded.

2) The actual response data were given as mean and standard deviation, together with the number of replicates for each of the measurement types.

3) A binary variable repeated measures was used to state if the studies contained repeated measures of the variable of interest, while the variable source stated the location of the data in the published paper (e.g., Fig3 or Table 2). The variable note_resp served to save notes on the data.

Step 7:  

Biotic covariates of response and manipulated taxa

For the taxa involved in the studies, either as response or as manipulated organisms, we recorded the following information:

1) The common name and phylogenetic information (including kingdom, phylum, class, order, family, genus and species name).

2) The occurrence of the organism as response or manipulated taxa.

3) The life-stage and information on biogeographic status.

4) Other covariates of species and functional groups, such as metabolism type, trophic status (e.g., autotroph, herbivore, carnivore, parasite), functional form (e.g., tree, shrub, periphyton, invertebrate, vertebrate), average generation time, body size, C:N:P ratios, etc.

Data Set Usage Rights

Obtain permission from data set owner(s)
Access Control:
Auth System:knb
Metadata download: Ecological Metadata Language (EML) File