Area. Predictions to high?

107 views
Skip to first unread message

Ias S

unread,
Aug 22, 2016, 4:09:00 AM8/22/16
to unmarked
Hi,

I have a model that appears to work very nicely ( according to chat values and bootstrap analysis).
I wanted to offset the area of my transect line.
When I used predict my values are a lot high then would be expected for this species.
Could I have made an error in my code while offsetting the transect size?

Thank you. Any help would be appreciated.

Ian

umf <- unmarkedFrameGDS(y=as.matrix(yDat), siteCovs=data.frame(sc.s,covs2), survey="line",  dist.breaks=db, numPrimary=1, tlength=covs2$translengt,  unitsIn="m")
head(umf)


> head(umf)
Data frame representation of unmarkedFrame object.
   y.1 y.2 y.3 y.4 y.5      ndvi      temp     prec       ele        riv       set       lake Activefire   firefreq   Firecomb
1    2   4   2   0   0 1.0772732 -1.428142 1.008960 0.4001723 -0.7200181 -1.072631 -0.8688888 -0.3098932 -0.3881905 -0.4593782
2    4   3   1   2   0 1.0772732 -1.428142 1.008960 0.4001723 -0.7200181 -1.072631 -0.8688888 -0.3098932 -0.3881905 -0.4593782
3    2   2   1   0   0 1.5158589 -1.428142 1.008960 0.3865088 -0.7078094 -1.021457 -0.8739560 -0.3098932 -0.3881905 -0.4593782
4    2   1   0   1   0 0.6304119 -1.428142 1.027776 0.3796771 -0.7868838 -1.087620 -0.8651899 -0.3098932 -0.3881905 -0.4593782
5    0   1   3   0   0 1.0027962 -1.428142 1.027776 0.4001723 -0.7868838 -1.072631 -0.8846160 -0.3098932  1.5244906  1.0652374
6    2   1   1   2   0 1.4993087 -1.428142 1.008960 0.3933406 -0.7868838 -1.042892 -0.9038299 -0.3098932 -0.3881905 -0.4593782
7    0   2   0   0   0 0.5890358 -1.428142 1.008960 0.4001723 -0.7868838 -1.087620 -0.8402437 -0.3098932 -0.3881905 -0.4593782
8    3   4   1   1   0 0.7793656 -1.428142 1.008960 0.3933406 -0.8234539 -1.051434 -0.8210714 -0.3098932 -0.3881905 -0.4593782
9    4   4   1   0   0 0.7793656 -1.428142 1.008960 0.3933406 -0.8234539 -1.051434 -0.8210714 -0.3098932 -0.3881905 -0.4593782
10   3   1   2   2   0 0.7959160 -1.428142 1.008960 0.4138358 -0.8234539 -1.042892 -0.8144356 -0.3098932 -0.3881905 -0.4593782
   uniqueID   OBS month translengt landcover area
1  20091A01 MULTI     5        500        10   25
2  20091A02 MULTI     5        500        10   25
3  20091A03 MULTI     5        500        30   25
4  20091A04 MULTI     5        500        30   25
5  20091A05 MULTI     5        500        20   25
6  20091A06 MULTI     5        500        30   25
7  20091A07 MULTI     5        500        30   25
8  20091A08 MULTI     5        500        30   25
9  20091A09 MULTI     5        500        30   25
10 20091A10 MULTI     5        500        30   25



# Load the raster layers of the predictors for the entire landscape.
install.packages(c('raster', 'rgdal', 'dismo', 'rJava'))
library(doParallel)
library(raster)
library(rgdal)

setwd("E:/paper 1/ANALYSIS/GIS and bird excel/GIS_projection UTM zone 42")



WWL<-gdistsamp(lambdaformula = ~riv + lake + prec + temp +offset(log(area)), phiformula = ~1,
    pformula = ~OBS, data = umf, keyfun = "exp", output = "abund",
    unitsOut = "ha", mixture = "NB")

summary(WWL)

layers <- list.files(pattern="tif$", full.names=FALSE)
layers

#stackkazc10a_Clip=river
#stackkazc12a_Clip=lake
#stackkazc2a_clipNDVI=
#stackkazc4a_Clip=temp
#stackkazc6a_Clip=prec
#stackkazc8a_Clip=elevation

#create four raster objects (coordinate system already specified).

riv <- raster("stackkazc10a_UTM42R1_Project1.tif")
lake<- raster("stackkazc12a_UTM42R1.tif")
prec<- raster("stackkazc6a_UTM42clip1.tif")
temp<- raster("stackkazc4a_UTM42clip1.tif")
ndvi<- raster("stackkazc2a_NDVIUTM42R1.tif")
ele<- raster("stackkazc8a_UTM42R1.tif")
firefreq<- raster("firefreq2000_15R.tif")


extent(temp)

###EXAMPLE RASTER
class       : RasterLayer
dimensions  : 2547, 1593, 4057371  (nrow, ncol, ncell)
resolution  : 856.3835, 856.3835  (x, y)
extent      : -542838.3, 821380.6, 4871142, 7052351  (xmin, xmax, ymin, ymax)
coord. ref. : +proj=utm +zone=42 +datum=WGS84 +units=m +no_defs +ellps=WGS84 +towgs84=0,0,0
data source : in memory
names       : stackkazc10a_UTM42R1_Project1
values      : -11.60904, 139.2227  (min, max)



#riv <- raster("stackkazc10a_Clip.tif")
#lake<- raster("stackkazc12a_Clip.tif")
#prec<- raster("stackkazc6a_Clip.tif")
#temp<- raster("stackkazc4a_Clip.tif")

sc.s # data set for the scaled variables
attributes (sc.s)

#$`scaled:center`
#       ndvi        temp        prec         ele         riv         set        lake  Activefire    firefreq    Firecomb
#107.7638431  58.2120523 205.3780557 253.4246731   0.1981767   0.2733345   0.4512685   0.1569073   0.2029562   0.3013076

#$`scaled:scale`
#       ndvi        temp        prec         ele         riv         set        lake  Activefire    firefreq    Firecomb
# 24.1685758  19.7543778  53.1457608 146.3752595   0.2406652   0.2432223   0.4376988   0.5063272   0.5228263   0.6559030

#Since we standardized the covariates during the modelling process, we need to transform the country-wide data using the same values

riv.s <- (riv-0.1981767)/0.2406652
lake.s <- (lake-0.4512685)/0.4376988
prec.s <- (prec-205.3780557)/53.1457608
temp.s <- (temp-58.2120523)/19.7543778


habitat <- stack(prec.s, temp.s, lake.s, riv.s)
                
               
names(habitat) <- c("prec", "temp", "lake", "riv")
habitat.s<-habitat
habitat.s$area<-(856.3835*856.3835)/10000 #cell resolution (856.3835)

#plot variables in model
plot(habitat.s, col=terrain.colors(100))


#predict species abundance per hectre
#WWL.density <- predict(WWL, type="lambda", newdata=habitat.s)
plot(WWL.density, axes=FALSE, col=terrain.colors(100))

Richard Chandler

unread,
Aug 22, 2016, 6:20:06 AM8/22/16
to Unmarked package
Hi Ian,

Can you show the estimates?

Richard


--
You received this message because you are subscribed to the Google Groups "unmarked" group.
To unsubscribe from this group and stop receiving emails from it, send an email to unmarked+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.



--
Richard Chandler
Assistant Professor
Warnell School of Forestry and Natural Resources
University of Georgia
Reply all
Reply to author
Forward
0 new messages