decompTSbfast: Decompose time series into trend, seasonality and remainder:...

In RETURN-project/BenchmarkRecovery: Benchmark recovery indicators derived from remote sensing time series

Description

Decompose time series into trend, seasonality and remainder: This function decomposes time series into three components using BFAST01 functionality: trend, seasonality and remainder. Trends are fitted using linear regression without breaks, seasonality is fitted using a first order harmonic function and the remainder equals the anomalies (i.e. time series - trend - seasonality).

Usage

decompTSbfast(df, nyr, nobsYr)

Arguments

df	a dataframe with time series that need to be decomposed. The dataframe needs to be structured as follows: each row represents a sampled pixel. The first two columns contain the latitude and longitude of the pixel. The next columns contain the time series values for each observation date.
nyr	number of years of the input time series
nobsYr	number of observations per year of the input time series

Value

a list containing the estimated seasonality, remainder, trend and seasonality coefficients. The seasonality is a dataframe with the seasonality of each pixel. Each row represents a sampled pixel. The first two columns contain the latitude and longitude of the pixel. The next columns contain the seasonality values for each observation date. The trend and remainder are dataframes with the trend and remainder of each pixel (dataframe is structured in the same way as the seasonality). Seasonality_coefficients is a dataframe with the coeficients of the fitted harmonic function. Each row represents a sampled pixel. The first two columns contain the latitude and longitude of the pixel. The next columns contain the coefficients of the fitted harmonic function.

RETURN-project/BenchmarkRecovery documentation built on July 13, 2021, 5:47 p.m.