duplisea / size Goto Github PK

library(devtools)
install_github("duplisea/size")

Load the package

library(size)

map.f()
points(-decdeg.f(ngsl.set$lon_deb*100),decdeg.f(ngsl.set$lat_deb*100),pch=".",col="red")

Use length cutoffs so catches outside of individuals between 10 cm and 100 cm is used.

LFS= CLF.f("all")
plot(LFS, 10, 100)

These are kind of boring plots without a lot of information for analysis but they are more diagnostics. They do act as a quick visualisation of the data and you can spot outliers quickly and figure out if they are real and further investigate them. You should not be trying to force a straight line through data that are clearly not linear.

LFS.lm= CLF.lm.fit(LFS, 10, 100)
plot(LFS.lm)

Fit linear models to the distribution each year and then plot the slopes and run a gam trend line through them. The interpretation on these kinds of plots is that the more negative the slope is, the fewer large fish there are relative to smaller ones. So for example fisheries which target large individuals may be reducing their abundance to such a degree that there is a paucity of individuals of large sizes over the whole community. It raises the idea of whether there is a slope which is ideal for a particular system and invokes ideas related to balanced fishing strategies, i.e. should we be targetting individuals over a large range of sizes so that the system is not “out of balance” given the size dependence of many physiological processes of individuals and ecological processes in marine communities.

Numbers per haul in log2 weight categories

tmp= SS.f("all",codeqc=792)
tmp$labundance= log2(tmp$abundance)
tmp$lbiomass= log2(tmp$biomass)
tmp= tmp[tmp$lg2W>3 & tmp$lg2W<16,]

par(mfcol=c(6,5),mar=c(1,1,1,1),omi=c(.5,.5,.01,.01))
years= sort(unique(tmp$year))
for (y in years){
  tmp2= tmp[tmp$year==y,]
  plot(tmp2$lg2W,tmp2$labundance,type="p",,ylim=c(0,max(tmp$labundance)), xlab="", ylab="", pch=20)
  nss.quad= lm(labundance~ poly(lg2W,2),data=tmp2)
  lines(tmp2$lg2W,predict(nss.quad),lwd=2)
  legend("topleft",legend=y,cex=0.6,bty="n",text.col="blue")
}
mtext(side=1,text="Log2 Weight-class",outer=T,line=2)
mtext(side=2,text="Abundance (no per tow)",outer=T,line=2)

Given the predator size distribution in the system in a year, the average predator/prey weight ratio and the variance (cv) of this this, the exponent relating consumption rate to body weight for a predator, the predation size spectrum can be determined. This is like a predation risk or the sizes of prey that would be most likely targetted by predators in the system in a year.

type ?PSS.f to see why I suppressed the warnings and I do not want them to appear in this markdown file.

pss= suppressWarnings(PSS.f("all",PPWR=10))
pss.scaled= pss$pred.risk/max(pss$pred.risk)
prey.size=10
plot(pss$year, pss.scaled[prey.size,],type="l",lwd=2,   ylim=c(0,max(pss.scaled[prey.size,])),col="red",xlab="Year",ylab="Predation risk for 10 cm prey")

pss= suppressWarnings(PSS.f("all",PPWR=100))
pss.scaled= pss$pred.risk/max(pss$pred.risk)
lines(pss$year, pss.scaled[prey.size,],lwd=2,col="blue")

pss= suppressWarnings(PSS.f("all",PPWR=30))
pss.scaled= pss$pred.risk/max(pss$pred.risk)
lines(pss$year, pss.scaled[prey.size,],lwd=2,col="black")

legend("topleft",bty="n",lwd=2,col=c("black","red","blue"),lty=1,legend=c("PPWR=30","PPWR=10","PPWR=100"),cex=0.8)

What you see is essentially the impact of the huge redfish year-class from 2011 becoming predators. if the PPWR is 10 then this year class starts having a large impacts on 10 cm prey by 2018 but if the PPWR is 30, prey of 10 cm are still mostly too large for this year class of redfish. The increase in predation risk will come further along. If the PPWR is 100, then there the predation index is much less dynamic. This is because the predators are not only quite different in size than their prey but the spread of prey sizes they target is also larger and therefore the predation impact of any particular predator cohort is much more diffuse over length classes.

The PLF for all species with a 30 cm threshold between small and large individuals

PLF.all= PLF.f(species.group="all",cutoff=30)

The PLF for all species with a 30 cm threshold between small and large individuals

PLF.dem= PLF.f(species.group="demersal",cutoff=30)

The PLF for all species with a 30 cm threshold between small and large individuals

PLF.gf= PLF.f(species.group="groundfish",cutoff=30)

The PLF for all species with a 30 cm threshold between small and large individuals

PLF.rf= PLF.f(species.group="code",codeqc=792,cutoff=20)

Plot the various PLF community indicators

ymax=max(c(PLF.all$plf,PLF.dem$plf,PLF.gf$plf))
plot(PLF.all$year, PLF.all$plf,xlab="Year",ylab="Proportion of individuals >=30cm",type="l",lwd=2,
  ylim=c(0,ymax))
lines(PLF.dem$year,PLF.dem$plf,col="blue",lwd=2)
lines(PLF.gf$year,PLF.gf$plf,col="green",lwd=2)
legend("topright",lwd=2,col=c("black","blue","green"),legend=c("everything","demersal","groundfish"),bty="n",cex=0.6)

Duplisea, D.E. 2005. Running the gauntlet: the predation environment of small fish in the northern Gulf of St Lawrence, Canada. ICES Journal of Marine Science 62: 412-416

Duplisea, D.E. and Castonguay, M. 2006. Comparison and utility of different size-based metrics of fish communities for detecting fishery impacts. Canadian Journal of Fisheries and Aquatic Sciences 63: 810-820.

Greenstreet, S.P., Rogers, S.I., Rice, J.C., Piet, G.J., Guirey, E.J., Fraser, H.M. and Fryer, R.J. 2010. Development of the EcoQO for the North Sea fish community. ICES Journal of Marine Science 68: 1-11.

Hahm, W. and Langton, R. 1984. Prey selection based on predator/prey weight ratios for some Northwest Atlantic fish. Marine Ecology Progress Series 19: 1-5.