I am trying to build a model to predict cumulative spring nitrous oxide (N2O) emissions based off of soil characteristics and winter weather. There are five sites. Within each site, there are multiple soil types, but soil does not change between years. Soil types within a site share the same weather every year. Weather variables are different every year.
The following variables are closely related and collinear within groups, (e.g. theta_s values depend upon how much sand/silt/clay are in the soil).
sand, silt, clay, theta_r, theta_s, silt_f_psa, silt_c_psa, Ks,
ph_h2o, cec7, cec82, bs7, ex_ca, ex_mg, ex_k,
oc, estimated_om, n_tot, estimated_c_to_n, ex_k_saturation
cold_sum, sum_tmax, sum_rad, snow_sum, precip_sum, cfd