Metadata for NL_muskrat_2013-2016.csv
Emiel van Loon, Daan Bos, Erik Klop & Ron Ydenberg
October 2019
e.e.vanloon@uva.nl
This data set supports the publication 'A Large-Scale Experiment to Evaluate the Effects of Trapping to Control Muskrats (Ondatra zibethicus) in The Netherlands' by Daan Bos, Emiel van Loon, Erik Klop and Ron Ydenberg. (the paper was accepted for publication in Wildlife Society Bulletin in 2020)
The Muskrat is an invasive species in Europe and in the Netherlands muskrat burrowing can compromise the integrity of dykes and hence poses a public safety threat. For that reason a control programme has been in effect since the arrival of the species in 1941. To investigate the relation between catch and effort and enhance prediction models, a large randomized controlled experiment was designed and conducted from 2013 till 2016. The publication by Bos et al. (2020) analyses the experimental results and here we present and document the experimental data.
The experiment ran for 39 months (December 2012 - March 2016) in the Netherlands.
Two treatments were applied:
- a manipulation of effort, relative that in a reference period (the 12 months preceding the start of the experiment in December 2012): 30% decrease, no change and 30% increase (3 levels)
- a manipulation of timing in which effort was spent: year-round (yr) or concentrated in winter and spring at the cost of time spent in summer (sc) (2 levels)
The treatments were applied to experimental units of 5 by 5 km (atlas squares), units which coincide with the national coordinate system of the Netherlands (EPSG:28992, also RD New; http://www.epsg-registry.org/). The Dutch mainland contains 2202 atlas squares in total. From these 117 atlas squares were selected for the experiment through a stratified random procedure. Table 1 below summarizes the distribution of the experimental units over the three strata (areas of high, medium or low muskrat suitability), as well as the two treatments. Table 2 characterises the three strata through the available muskrat habitat (length of water ways) and the historical catch and effort data.
Table 1. Distribution of experimental units
(117 atlas squares in total) over the
two treatments within each of the
pre-defined strata.
=======================================
decrease control increase <-- effort treatment
Stratum -------- ------- --------
yr sc yr yr sc <-- time treatment
--- --- --- --- ---
low 7 6 13 7 6
medium 7 6 13 7 6
high 7 6 13 7 6
=======================================
Table 2. Characterisation of the 3 strata through the
average length of the water ways per atlas
square as well as the average catch (numbers
of muskrat caught) and effort (field hours),
both per atlas square per year.
The column labeled N gives the number of
atlas squares in the Netherlands within the
respective stratum.
=====================================================
average over 2009-2011
-----------------------
stratum N waterway (km) catch (nr) effort (hr)
------- --- ------------- ---------- -----------
low 701 93 27 145
medium 460 285 124 480
high 76 506 796 1730
=====================================================
For each atlas square we aggregated the records by 'season': spring (day-of-year 57 - 140); summer (141 - 224); autumn (225 - 308); and winter (309 - 56 in next year). The experiment ran for 13 successive seasons, identified in the analysis by the variable named 'timenum', from winter 2012/2013 (timenum = 1) through winter 2015/2016 (timenum = 13).
Each record in the data set refers to the observations concerning one atlas square during one season. Hence there are 117 (experimental atlas squares) * 13 (seasons) = 1521 records in the data set with experimental results.
In the experimental period, muskrat trapping continued according to standard procedures in the remainder of the Netherlands, covering 1401 non-experimental atlas squares (these non-experimental atlas squares are all areas in the Netherlands were effort was spent from 2012 to 2016 to catch muskrats). Also the data for these non-experimental atlas squares data for the 13 seasons seasons is included in the dtaa set (1401*13 = 18213 records).
The data matrix contains 19734 rows and 17 columns. Each row contains information for one atlas square in one season. Each column represents a variable. The labels, definitions and data types of the variables are as follows.
-
as = atlas square, a unique identifyer for a 5 km by 5 km block fitting in the dutch coordinate system (integer)
-
rwa = name of regional water authority, 21 labels in total (character)
-
year = year of observation (character)
-
yearnum = year of observation (integer)
-
season = season: winter, spring, summer, autumn (character)
-
seasonnum = season: 1 to 4 (integer)
-
time = incremental time since start of experiment z-transformed values of timenr (float)
-
timenum = incremental time since start of experiment: 1-13 (integer)
-
stratum = the stratum in which the as falls: low, medium and high suitability for muskrats; NA values for all atlas squares that are not part of the experiment (character)
-
eff = the effort treatment: control, decrease, increase and no experiment (character)
-
temp = the temporal treatment: the effort to catch muskrats is distributed year-round (yr) or concentrated in the spring and winter seasons (sc); NA values for all atlas squares that are not part of the experiment (character)
-
hist = average catch per unit effort in the 3-year period 2010-2012 (float)
-
neighb = average catch per unit effort in the 8 atlas squares surrounding the focal atlas square (float)
-
soil = soil type (sand, loam, clay, peat), built-up area or water (character)
-
catch = number of muskrats caught (integer)
-
effort = number of hours spent on muskrat trapping (float)
-
cpue = catch per unit effort, i.e. catch/effort (float)
Missing data in the data matrix is encoded by the text label NA
The file "NL_muskrat_2013-2016.csv" uses the comma as the field separator and the point as a decimal sign. It has a header with the variable labels at the first row. The variable names in the header as well as character data are between double qoutes (".."), the missing data label (NA) is not in quotes.
This data set is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
