Integration　of　large-area,　economically　driven　macro-models　and　small-area,　biophysically　based　models　in　the　Canadian　agricultural　sector　was　described　recently　in　the　development　of　a　land-use　allocation　model　(LUAM).　We　have　since　developed　and　integrated　an　improved　methodology　for　allocating　crop　area　changes　generated　for　large　areas　by　the　Canadian　Regional　Agricultural　Model　(CRAM)　to　much　smaller　Soil　Landscapes　of　Canada　(SLC)　polygons.　Validation　of　outputs　showed　considerable　improvement.　The　new　coefficients　of　determination　(R-2)　between　simulated　and　actual　data,　with　previous　values　in　brackets,　were　0.69　for　fodder　corn　(0.54),　0.88　for　wheat　(0.62),　0.77　for　hay　(0.26),　0.54　for　alfalfa　(not　previously　reported),　0.88　for　soya　bean　(0.26)　and　0.86　for　grain　corn　(0.22).　The　best　result　was　obtained　for　soya　bean,　with　a　normalized　root　mean　square　error　(NRMSE)　of　0.31%,　and　the　poorest　for　alfalfa,　with　NRMSE　=　17.34%.