Population and community time series are often scaled (mean-centered and scaled by standard deviation) before analysis. While this works with linear models, it does generally not with nonlinear analysis. This also depends on whether you analyze on linear or on log-scale.
