phenofit: An R package for extracting vegetation phenology from time series remote sensing

Abstract

1. Satellite-derived vegetation indices (VIs) provide a way to analyse vegetation phenology over decades globally. However, these data are often contaminated by different kinds of optical noise (e.g. cloud, cloud shadow, snow, aerosol), making accurate phenology extraction challenging. 2. We present an open-source state-of-the-art R package called phenofit to extract vegetation phenological information from satellite-derived VIs. phenofit adopts state-of-the-art phenology extraction methods, such as a weight updating function for reducing optical noise contamination, a growing season division function for separating the VI time series into different vegetation cycles, and rough and fine fitting functions for reconstructing VI time series. They work together to make phenology extraction from frequently contaminated VIs easier and more accurate.3. Compared against other widely used phenology extraction tools, for example, TIMESAT and phenopix, phenofit provides flexible input and output options, a practical growing season division function, rich curve fitting and phenology extraction functions, and robust performance under different kinds of optical noise.4. In addition to working with VIs from mesoscale satellites (e.g. MODIS and AVHRR), phenofit can also reconstruct vegetation time series and extract phenology using other sources, such as VIs from high-resolution optical satellites (e.g. Sentinel-2 and Landsat) and radar satellites (e.g. Sentinel-1), vegetation greenness indices from digital cameras and gross primary production estimations from eddy-covariance sites. As such, phenofit can contribute to the study of ecological process dynamics and assist in effective modelling of global change impacts on vegetation, as caused by climate variability and human intervention. Code and | 1509 Methods in Ecology and Evoluঞon KONG et al.

‘LandsatTS': an R package to facilitate retrieval, cleaning, cross‐calibration, and phenological modeling of Landsat time series data

“…Users interested in other aspects of vegetation phenology (e.g. timing of spring onset or fall senescence) could extract and process Landsat data using ‘LandsatTS', but then capitalize on tools provided by other R packages (http://www.r-project.org), such as the new ‘phenofit' package that provides state‐of‐the‐art tools for fitting phenological models (Kong et al 2022). Alternatively, users who are interested in phenological modeling with other data sources (e.g.…”

Estimating phenology and phenological shifts with hierarchical modeling

“…The majority of tools in R for phenological modeling involve remote sensing data. Examples include (1) the phenor R package for linking phenological observations from remote sensing data and climate drivers in North America (Hufkens et al, 2018), the (2) phenofit package for extracting phenology information from remotely sensed vegetation data (Kong et al, 2021), and (3) the CropPhenology package for the analysis of data from agriculture (Araya et al, 2018). Several existing packages also include statistical approaches to fitting observed phenological data to parametric distributions; for example, the package nlstimedist (Steer et al, 2019) fits an extension of the exponential distribution to phenological data, and the phenology package fits parametric curves to phenological data, though without heavy‐tailed distributions or random variation (Girondot, 2020).…”

Experimental warming causes mismatches in alpine plant-microbe-fauna phenology

“…Their 95 and 5% quantiles were treated as the upper and lower limits of NDVI sequences to constrain mean daily NDVI values in a reasonable range 51 . We finally used the Savitzky–Golay filter in the R “phenofit” package to filter and denoise NDVI sequences 53 . We used a double logistic (DL) function to obtain the smooth seasonal dynamic curves of NDVI, and to interpolate the missing values 54 .…”