Incorporating spatial autocorrelation in rarefaction methods: implications for ecologists and conservation biologists