是否有使用 Spearman 方法计算栅格 p 值的函数？

Question

I can not succeed to calculate the p value for the Spearman correlation for a series of the few rasters. I created a code to make spearman correlation. It works, but if I want to show me p value, using cor.test (), I have a error.

A reproducible example using a raster with 6 layers to compute the correlation between the cells of layers 1,2,3 and 4,5,6.

library(raster)
set.seed(89)
z <- brick(nrow=10, ncol=10, nl=6)
values(z) <- runif(600)

### Spearman correlation 
r <- calc(z, fun=function(x) cor(x[1:3], x[4:6], method='spearman'))

### p-value -- here I have the error
r.pvalue <- calc(z, fun=function(x) cor.test(x[1:3], x[4:6], method='spearman'))
#Error in setValues(out, x) : 
#  values must be numeric, integer, logical or factor

Answer 1

The function used in calc must return a number (or numbers). cor returns a number, but cor.test does not.

cor(1:3, 3:1)
#[1] -1

x <- cor.test(1:3, 3:1)
class(x)
#[1] "htest"
x
#        Pearson's product-moment correlation
#
#data:  1:3 and 3:1
#t = -Inf, df = 1, p-value < 2.2e-16
#alternative hypothesis: true correlation is not equal to 0
#sample estimates:
#cor 
# -1 

So we need to extract the p value from the object returned by cor.test . If you look at ?cor.test or inspect str(x) you can see that you can use

cor.test(1:3, 3:1)$p.value
#[1] 0

And thus you can do

r.pvalue <- calc(z, fun=function(x) cor.test(x[1:3], x[4:6], method='spearman')$p.value)