Initiating the model in 1959 with the observed presences up until 1960. Similar to the model initiated in 1939, the spread is initially slower than observed in the full model, but prediction accuracy increases in 1980 because the location of populations in 1940 is sufficient to explain occurrence in 1980. It seems, however that there may have been greater spread in 1980 than shown in this figure because results in 2000 suffer from under-prediction.
