S. K. Morgan Ernest, Thomas J. Valone, and James H. Brown. 2009. Long-term monitoring and experimental manipulation of a Chihuahuan Desert ecosystem near Portal, Arizona, USA. Ecology 90:1708.

RODENT METADATA

INTRODUCTION

The long-term data for the rodent community at the Portal Project has been used to address a variety of questions including (1) the effects competition from kangaroo rats on other granivorous rodents, (2) responses of rodents to climatic variability, and (3) the long-term stability and dynamics of a desert rodent community. This metadata file describes the field data collection techniques and data structure for the rodent data set. Further details on the rodent community and data collection are explained in this file.

METADATA CLASS I. DATA SET DESCRIPTORS

A. Data set identity:

Title: Long-term monitoring and experimental manipulation of a desert rodent community in the Chihuahuan Desert near Portal, Arizona, USA.

B. Data set identification code: Portal_rodents_19772002.csv

C. Data set description

Principal Investigators:

S. K. Morgan Ernest, Department of Biology, Utah State University, Logan UT 84322

Thomas J. Valone, Department of Biology, Saint Louis University, Saint Louis MO 63103

James H. Brown, Department of Biology, University of New Mexico, Albuquerque NM 87131

Abstract: The data set covers a 25 year period (1977–2002) of monthly rodent surveys. Each year rodents are trapped on all 24 experimental plots and information on each captured rodent is contained in this file. The data set should prove useful for studying population dynamics and species interactions. Sampling is ongoing and data will be added over time. 

D. Key words:  desert rodents, arid ecosystem, climate, competition, LTREB data.

CLASS II. RESEARCH ORIGIN DESCRIPTORS

A. Overall project description

Identity: Capture data for every desert rodent caught on the 20 ha study area in the Chihuahuan Desert near Portal, Arizona, USA.

Originators: Drs. James H. Brown, Diane W. Davidson, and James Reichman

Period of Study: 1977–2002

Objectives: To monitor the community- and population-level dynamics of desert rodents and competitive interactions among rodent species.

Source(s) of funding: See metadata.htm

B. Specific subproject description

1. Site description: Within the 20 ha study area there are 24 experimental plots. Each plot has an area of 0.25 ha and is fenced to regulate rodent access to the plot. Access is regulated using gates cut into the fencing. Large gates (3.7 × 5.7 cm) allow all small mammals to access plots. Small gates (1.9 × 1.9 cm) exclude kangaroo rats (Dipodomys) whose inflated auditory bullae make their skulls too large to pass through the gates. Rodent removal plots do not contain any gates and animals caught on those plots are removed and released outside the cattle exclosure fence.

On each plot there are 49 permanent trapping stations marked by rebar stakes forming a 7 × 7 grid. Every stake on a plot has a unique identifying number denoting the coordinate of that stake on that plot. For example, stake 11 is the first stake on the first row. Rows are numbered 1 through 7 going from the most northern row to the most southern. Columns are numbered 1 through 7 going from the most western column to the most eastern (See metadata.htm, Fig. 1).

Treatments: See metadata.htm, Table 1, for details on treatment assignments for each plot.

Data Collection Period, Frequency: From 1977–2002, plots were trapped around each new moon – which occurs approximately once a month, though occasionally blue moons do occur, resulting in 2 separate surveys in a month. Occasionally, months are missed. Months that are entirely missed are not noted in the database. Sometimes weather or other unforeseen occurrences prevent the complete trapping of a survey, these are noted in the database (see Table 1 below).

3. Research Methods

Field: The site is surveyed for rodents approximately once each month. The survey occurs as close to the new moon as possible to minimize external effects on trapping success which could be misconstrued as actual changes in populations. During a survey, each plot is trapped for one night with treatments divided evenly between nights to eliminate differences between controls and treatments caused by environmental differences on different nights. When a plot is surveyed, all gates are closed to ensure that only resident individuals are captured. At each stake, one Sherman live-trap is placed and baited with millet seed. Traps are collected the next morning and individuals processed. Each individual was tagged and data on species, location caught (plot and stake), sex, reproductive condition, weight, and hindfoot length were recorded. Until 1993, individuals were tagged using either ear tags or toe tags. After 1993, individuals were predominately tagged using Passive Integrated Transponder (PIT) Tags. Use of PIT tags was halted from 1997–2000 while we waited for a patent infringement lawsuit to be resolved against the company that supplied our tags and scanning system (Troyvan). In 2000 we resumed using PIT tags, using a different system (Biomark). It is noted in the database if either an animal escaped before all information was gathered, was removed from the site because it was caught on a plot it was supposed to be excluded from, or died during trapping.

Most data has been consistently collected over the timespan of the study. The one exception is that initially data were also collected on the location of the burrow for Dipodomys spectabilis and Neotoma albigula. This was discontinued after 1994.

Taxonomy and systematics: Taxonomy and identification of species is consistent with Mammals of Arizona by Hoffmeister (1986), with the exception that some species identified as Perognathus in Hoffmeister (1986) are now a separate genus (Chaetodipus).

4. Project personnel:  Our current estimate of the number of people who have assisted with the rodent survey is currently well over 100. Many of these people assisted on a volunteer basis and this work would not have been possible without their help.

CLASS III. DATA SET STATUS AND ACCESSIBILITY

A. Status

Latest Update: 15 February 2011

Latest Archive date: 15 February 2011

Metadata status: The metadata are complete and up to date.

Data verification: In general, quality control activities are on-going. After every survey, every record for a captured individual from that survey is compared to preexisting records for that tag number to assure that errors have not been made in species or sex identification in the field. For example if a female Dipodomys merriami with tag 0001 was captured, previous incidences of that tag number are compared to make sure that it has always been identified as a female Dipodomys merriami. Discrepancies are settled using either additional information (i.e., if an individual has been recorded as a non-reproductive female and is now a scrotal male – a trait that is hard to misidentify - the male designation overrides the previous information) or by using the value recorded in the majority of identifications if more than one previous record exists. Because tags numbers are recycled over the course of 30 years these comparisons are restricted to the previous 3 years which is usually less than the tag number recycle rate and approaches the maximum number of years a rodent would be expected to live.

Data have also been periodically spot checked by people using the database who find oddities or outliers. Questions regarding particular records are answered by referring to the original datasheets. Datasheets for the rodent database currently reside with S. K. Morgan Ernest at Utah State University.

For the publication of the data set, an additional and extensive quality control effort was initiated. Information outside the norms (e.g., stake numbers that do not exist, undocumented species codes, plots with no animals recorded, weights either too small or too large for the identified species) was systematically checked and compared to original datasheets. Corrections were made based on datasheets or notes were added to the database if correction was not possible. A list of notes and their meaning is included below (Table 1). In addition, random spot checking of records also occurred during this process.

B. Accessibility

Storage location and medium: (Ecological Society of America data archives [Ecological Archives], URL published in each issue of its journals). Original data files reside with S. K. Morgan Ernest on two separate machines. Original datasheets also reside with S. K. Morgan Ernest.

Contact person: S. K. Morgan Ernest, Department of Biology, Utah State University, Logan UT 84322, Phone: 435-797-8751 E-mail: [email protected]

Copyright restrictions: None.

Proprietary restrictions: None. However, we do request that authors of publications using the rodent database notify S. K. Morgan Ernest (contact information above) of publication of their study. This helps us by allowing us to make accurate reports to the National Science Foundation and document that the scientific community finds the data from this study to be useful.

Costs: None.

CLASS IV. DATA STRUCTURAL DESCRIPTORS

A. Data Set File

Identity: Portal_rodents_19772002.csv

Size: 35,550 rows (including header), 2.10 Mb.

Format and storage mode: ASCII text, comma delimited. No compression scheme used.

Header information: The first row of the file contains the variable names below.

Alphanumeric attributes: Mixed.

B. Variable information

Table 1. Column information for Portal_rodents_19772002.csv
Each row in this data set is the information collected on an individual rodent during that survey. If an individual is caught multiple times, it has multiple entries in the data set. For most variables, a blank cell indicates missing data. The exceptions are as follows: reproductive variables (Reprod, Testes, Vagina, Pregnant, Nipples, Lactation) and notes (note1, note2, note3, note4, note5). For reproductive variables, blanks can indicate either missing data or lack of reproductive condition. If the variable ‘Reprod’ = Z then the animal is not in reproductive condition and blanks in the other reproductive condition variable indicate that the individual did not display any of those conditions. If ‘Reprod’ = blank, but at least one other reproductive value is recorded, then blank cells again indicate the individual did not display any of those conditions which are blank. If ‘Reprod’ = blank and all other reproductive variables are also blank, then reproductive information for was not collected for that individual. For notes, blanks do not denote missing information. Instead, they denote nothing relevant to report.

Table 2. SPECIES CODES in Portal_rodents_19772002.csv
Since we record all animals caught in traps, some entries are for non-rodents.

CLASS V. SUPPLEMENTAL DESCRIPTORS

A. Publications using the data set: See metadata.htm

LITERATURE CITED

Hoffmeister, D. F. 1986. Mammals of Arizona. University of Arizona Press, Tucson, Arizona, USA.