Parasite Infections in Cod

Description

Data on parasite infection in cod along the coast of Finmark.

Usage

data("CodParasites")

Format

A data frame containing 1254 observations on 10 variables.

intensity

Number of parasites.

prevalence

Factor indicating presence of parasites (i.e., intensity > 0).

area

Factor indicating sampling area.

year

Factor indicating sampling year.

depth

Depth at which the fish were caught.

weight

Weight of the fish.

length

Length of the fish.

sex

Factor indicating sex of the fish.

stage

Factor indicating stage of the fish.

age

Age of the fish.

Details

The red king crab Paralithodes camtschaticus was deliberately introduced to the Barents Sea in the 1960s and 1970s from its native area in the North Pacific. The carapace of these crabs is used by the leech Johanssonia arctica to deposit its eggs. The leech in turn is a vector for the blood parasite Trypanosoma murmanensis that can infect marine fish, including cod.

Hemmingsen et al. (2005) examined cod for trypanosome infections during annual cruises along the coast of Finnmark in North Norway over three successive years and in four different areas (A1 Sørøya; A2 Magerøya; A3 Tanafjord; A4 Varangerfjord). They show that trypanosome infections are strongest in the area Varangerfjord where the density of of red king crabs is highest. Thus, there is evidence that the introduction of the foreign red king crabs had an indirect detrimental effect on the health of the native cod population. This situation stands out because it is not an introduced parasite that is dangerous for a native host, but rather an introduced host that promotes transmission of two endemic parasites.

Zuur et al. (2009) reanalyze the data using binary and count data regression models in Chapters 10.2.2, 11.3.2, 11.4.2, 11.5.2.

Source

The data are taken from the online supplements of Zuur et al. (2009). http://highstat.com/index.php/mixed-effects-models-and-extensions-in-ecology-with-r

References

Hemmingsen W, Jansen PA, MacKenzie K (2005). “Crabs, Leeches and Trypanosomes: An Unholy Trinity?”, Marine Pollution Bulletin 50(3), 336–339.

Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM (2009). Mixed Effects Models and Extensions in Ecology with R, Springer-Verlag, New York.

Examples

library("countreg")

## load data
data("CodParasites", package = "countreg")

## Table 1 from Hemmingsen et al. (2005)
## number of observations
xtabs(~ area + year, data = CodParasites)

               year
area            1999 2000 2001
  soroya         147   55   70
  mageroya        98   50  107
  tanafjord      183   75  157
  varangerfjord  139   50  123

## prevalence of parasites (NAs counted as "yes")
tab <- xtabs(~ area + year + factor(is.na(prevalence) | prevalence == "yes"),
  data = CodParasites)
round(100 * prop.table(tab, 1:2)[,,2], digits = 1)

               year
area            1999 2000 2001
  soroya        61.2 70.9 10.0
  mageroya      32.7 36.0 31.8
  tanafjord     33.9 57.3 28.7
  varangerfjord 75.5 88.0 65.9

## omit NAs in response
CodParasites <- subset(CodParasites, !is.na(intensity))

## exploratory displays for hurdle and counts
par(mfrow = c(2, 2))
plot(factor(intensity == 0) ~ interaction(year, area), data = CodParasites)
plot(factor(intensity == 0) ~ length, data = CodParasites, breaks = c(15, 3:8 * 10, 105))
plot(jitter(intensity) ~ interaction(year, area), data = CodParasites,
  subset = intensity > 0, log = "y")
plot(jitter(intensity) ~ length, data = CodParasites, subset = intensity > 0, log = "y")

## count data models
cp_p   <-    glm(intensity ~ length + area * year, data = CodParasites, family = poisson)
cp_nb  <- glm.nb(intensity ~ length + area * year, data = CodParasites)
cp_hp  <- hurdle(intensity ~ length + area * year, data = CodParasites, dist = "poisson")
cp_hnb <- hurdle(intensity ~ length + area * year, data = CodParasites, dist = "negbin")
AIC(cp_p, cp_nb, cp_hp, cp_hnb)

       df       AIC
cp_p   13 20377.862
cp_nb  14  5030.673
cp_hp  26 13687.576
cp_hnb 27  4937.085

BIC(cp_p, cp_nb, cp_hp, cp_hnb)

       df       BIC
cp_p   13 20443.935
cp_nb  14  5101.829
cp_hp  26 13819.722
cp_hnb 27  5074.314

## rootograms
if(require("topmodels")) {
par(mfrow = c(2, 2))
br <- 0:51 - 0.5
rootogram(cp_p, breaks = br, main = "Poisson")
rootogram(cp_nb, breaks = br, main = "Negative Binomial")
rootogram(cp_hp, breaks = br, main = "Hurdle Poisson")
rootogram(cp_hnb, breaks = br, main = "Hurdle Negative Binomial")
}