Spatial and temporal autocorrelations affect Taylor's law for US county populations: Descriptive and predictive...

Area of Science:

• Ecology
• Demography
• Statistical modeling

Background:

• Taylor's law quantifies population distribution using mean-variance relationships.
• Understanding population fluctuations is key in ecology and demography.

Purpose of the Study:

• To extend Taylor's law by incorporating temporal and spatial autocorrelations.
• To develop and test descriptive and predictive statistical models for Taylor's law.
• To apply these models to U.S. county population data (1790-2010).

Main Methods:

• Generalized least-squares models to account for autocorrelations.
• Analysis of three statistical models predicting Taylor's law form and slope.
• Application to U.S. decennial census data.

Main Results:

• Temporal and spatial autocorrelations significantly impact Taylor's law slope estimates.
• Generalized least-squares models outperform ordinary least-squares when accounting for autocorrelations.
• Combined effects of autocorrelations and demographic factors influence human population Taylor's law.

Conclusions:

• Careful evaluation of descriptive model assumptions is vital for interpreting Taylor's law.
• Accounting for temporal and spatial dependencies enhances the accuracy of population distribution analysis.
• The study provides a more robust framework for analyzing population dynamics.