PhenoGu: A function to extract phenological thresholds according to Gu...
In phenopix: Process Digital Images of a Vegetation Cover

Description Usage Arguments Details Value Note Author(s) References See Also Examples

A function to extract phenological thresholds according to Gu et al. 2009. This is a rather internal function. Use PhenoExtract with method='gu' instead.

1	PhenoGu(x, fit, uncert = FALSE, breaks, sf, ...)

`x`	A named vector with the parameters of the equation used to fit the data.
`fit`	A list structured as in output from the fitting procedures, such as GuFit, KlostermanFit, ElmoreFit, BeckFit.
`uncert`	Currently unused
`breaks`	Currently unused
`sf`	Scaling factors required to normalize the data prior to the fitting. If the function is called by e.g. `greenProcess` sf is automatically calculated.
`...`	Further arguments, currently not used

Threshold extraction is performed according to Gu et al (2009). Briefly, from the fitting equation (be it from Klosterman, Elmore, Beck or Gu fitting) and the correspondent parameters, the first derivative is extracted. Peak recovery rate (prr) is defined as the maximum of the first derivative and represent the maximum slope of the increasing part of the curve. Correspondingly peak senescence rate (psr) is the minimum of the first derivative. The recovery line and the senscence line are tangent to the fitting curve with slope equal to prr and psr respectively. Baseline and maxline are further defined as the horizontal lines corresponding to the minimum and the maximum of the curve. The intersection between recovery line and baseline defines the beginning of the growing season, i.e upturn date (UD). The intersection between recovery line and maxline defines the reacing of the plateau, i.e. stabilization date (SD). The intersection between senescence line and baseline defines the end of the growing season, i.e. recession date (RD). In the original implementation by Gu et al. the intersection between senescence line and maxline would define the downturn date (oldDD). To account for decreasing plateau typical of a number of seasonal trajectories, we have further defined a plateau line, which is a linear fit between SD and oldDD. The plateau line would roughly correspond to the maxline if the plateau is horizontal, whereas it will be a decreasing line in case of a decreasing plateau. Therefore, in our implementatio DD is defined as the intersection between the plateau line and the senescence line.

A named vector of length 9 with upturn date (UD), stabilizazion date (SD), downturn date (DD), recession date (RD), maximum of the fitting curve (maxline), minimum of the fitting curve (baseline), peack recovery rate (prr), peack senescence rate (psr) and the slope of the plateau line (plateau.slope).

Since this threshold extraction is based on a derivable function, it cannot be performed on raw data. Uncertainty estimation with this method on a fitted curve from SplineFit is currently not implemented. Instead you can use PhenoGu in a for loop cycling in the uncertainty dataframe columns.

Gianluca Filippa <gian.filippa@gmail.com>

Gu L, Post WM, Baldocchi D, Black TA, Suyker AE, Verma SB, Vesala T, Wofsy SC. (2009) Characterizing the Seasonal Dynamics of Plant Community Photosynthesis Across a Range of Vegetation Types. In: Phenology of Ecosystem Processes (Ed: Noormets A, Springer New York), pp 35-58.

PhenoExtract

data(bartlett2009.fitted)
gu.phenophases <- PhenoGu(x=bartlett2009.fitted$fit$params, 
	fit=bartlett2009.fitted$fit, sf=quantile(bartlett2009.fitted$fit$predicted, c(0.1, 0.9)))
plot(bartlett2009.fitted$fit$predicted)
abline(v=gu.phenophases[1:4], col=palette())
mtext(names(gu.phenophases[1:4]), at=gu.phenophases[1:4], 
	line=-2:-5, col=palette()[1:4])