106handout3 [Compatibility Mode]

1autocorrelation

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: : (1)N: number of values. (2) △t: time interval. Time series analysis: detecting, describing, and modeling climate variability and impacts. Care must be taken when you work on a time series data. Aliasing: The information that varies at shorter timescales appears as information at longer timescales. (1) Auto‐correlation Auto‐correlation: The simple linear correlation of a time series with its own past; that is, the correlation between a variable A sampled at time t and the same variable A sampled at a later time t+△t, t+2Δt, .... Example Auto‐correlation between U(t) and U(t+1) Index: 0 1 2 3 4 5 6 7 8 9 Time: noon 1 2 3 4 5 6 7 8 9 pm U(t): 5 6 5 4 7 5 3 5 4 6 U (t+1) 6 5 4 7 5 3 5 4 6 Example Auto‐correlation between U(t) and U(t+1) Index: 0 1 2 3 4 5 6 7 8 9 Time: noon 1 2 3 4 5 6 7 8 9 U(t): 5 6 5 4 7 5 3 5 4 6 U (t+1) 6 5 4 7 5 3 5 4 6 U(t+1) Scatter plot of hourly wind speed versus the following hour’s wind speed U(t) rAB COVAR(A,B) AB U(t+1) Best visualization of autocorrelation is by plotting the autocorrelation as a function of lag time, that is autocorrelation function (acf). 1 U(t) U(t): 5 6 5 4 7 5 3 5 4 6 U (t+1) 6 5 4 7 5 3 5 4 6 Best visualization of autocorrelation is by plotting the autocorrelation as a function of lag time, that is autocorrelation function (acf). U(t): 5 6 5 4 7 5 3 5 4 6 U (t+1) 6 5 4 7 5 3 5 4 6 U(T+2) 5 4 7 5 3 5 4 6 12 Best visualization of autocorrelation is by plotting the auto‐ correlation as a function of lag time, that is autocorrelation function (acf). U(t), U(t+1); U(t), U (t+2); U(t), U(T+3);….. U(t), U(t+6);…… U(t), U(t+24);……… Acf: show the amount of persistence in time series. All acfs exhibit a decay with lag time. The time that it needs to make acf near 0 indicates the memory. Best visualization of autocorrelation is by plotting the auto‐ correlation as a function of lag time, that is autocorrelation function (acf). U(t), U(t+1); U(t), U (t+2); U(t), U(T+3);….. U(t), U(t+6);…… U(t), U(t+24);……… From statistics, autocorrelation can be established as U(t)=f1* U(t-1) + f2* U(t-2) + … Best visualizat...
View Full Document

This document was uploaded on 01/22/2014.

Ask a homework question - tutors are online