[R-sig-ME] Correlations for spatio-temporal data ?

[David Duffy]

On Sun, 11 Dec 2011, Jian Zhang wrote:

>
> I have a dataset like this:
>
>      sites years Var1 Var2
>      1 1960 505 3.013833
        ....
>      5 1960 476 4.337246
>      5 1961 540 8.601403
>      5 1962 610 18.99233
>      5 1963 479 1.336739
>      5 1964 822 43.73037
>      5 1965 692 30.30235
>
> Meantime, I have the spatial location (longitude, latitude) for each 
> site. Among these sites, spatial autocorrelation exists. Within each 
> site, for the variables Var1 and Var2, temporal autocorrelation 
> also exists. I want to calculate the correlation between Var1 and Var2 
> using correlation coefficient. My question is that how can I add the 
> information of spatial and temporal autocorrelation into my analysis. I 
> know that, for only spatial autocorrelation, we can use modified 
> t-test to do the analysis. But I have no idea how to do it when 
> considering the spatial and temporal autocorrelation together?

If the data is as regular as in your example (fixed set of years for every 
site etc), you could fit this as a structural equation model to the 
summary covariance matrix using, for example, the sem package.

Just 2c, David Duffy.

-- 
| David Duffy (MBBS PhD)                                         ,-_|\
| email: davidD at qimr.edu.au  ph: INT+61+7+3362-0217 fax: -0101  /     *
| Epidemiology Unit, Queensland Institute of Medical Research   \_,-._/
| 300 Herston Rd, Brisbane, Queensland 4029, Australia  GPG 4D0B994A v