第一个代码trend=~poly(coords[,1],degree=2)

rl <- likfit(pre, cov.model = "mat",kap = 0.5,ini = c(10,2),
             trend=~poly(coords[,1],degree=2),lambda = 0.5, nug = 4)
rl

第二个代码trend=~coords[,1]+I(coords[,1]^2)

ml <- likfit(pre, cov.model = "mat",kap = 0.5,ini = c(10,2),
             trend=~coords[,1]+I(coords[,1]^2),lambda = 0.5, nug = 4)
ml

这两个趋势的意思应该是一样的
可是运行的结果却不一样，不知道为什么
第一个结果是

> rl
likfit: estimated model parameters:
   beta0    beta1    beta2    tausq  sigmasq      phi
"11.189" "30.754" "35.919" " 4.431" "11.079" " 2.000"
Practical Range with cor=0.05 for asymptotic range: 5.992074

likfit: maximised log-likelihood = -485.5
第二个结果是
> ml
likfit: estimated model parameters:
     beta0      beta1      beta2      tausq    sigmasq        phi
"216.2002" " -0.1422" "  0.0000" "  4.4312" " 11.0790" "  2.0002"
Practical Range with cor=0.05 for asymptotic range: 5.992056

likfit: maximised log-likelihood = -485.5
出现这种奇怪的现象