How to predict future land use

[qiangxyz]

We have land use data (data from the Plum Island) from 3 years, say 1985, 1991 and 1999, and we would like the predict the future land use map for 2005. We predict it with the following code:

dmd <- approxExtrapDemand(obs=obs, tout=14:20)
clues.rules <- matrix(data=1, nrow=3, ncol=3, byrow=TRUE) 
clues.parms <- list(jitter.f=0.0002,
                    scale.f=0.000001,
                    max.iter=1000,
                    max.diff=50, 
                    ave.diff=50)
clues.model <- CluesModel(obs=obs,
                          ef=ef,
                         models=glm.models,
                         time=14:20,
                         demand=dmd,
                         elas=c(0.2,0.2,0.2),
                         rules=clues.rules,
                         params=clues.parms)
clues.model <- allocate(clues.model)
clues.model

And we got: Does it means that the prediction for 2005(t=20) is based on map of 1985( t = 0)? Can we prediction the land use for 2005(t=20) based on map of 1999?