Assessing land surface phenology in Araucaria-Nothofagus forests in Chile with Landsat 8/Sentinel-2 time series

• Published in: International Journal of Applied Earth Observation and Geoinformation
• Main Authors: E. Kosczor, M. Forkel, J. Hernández, D. Kinalczyk, F. Pirotti, E. Kutchartt
• Format: Article in Journal/Newspaper
• Language: English
• Published: Elsevier 2022
• Subjects: Araucaria araucana; Land surface phenology; Plant phenology; Topography; Climate; Random Forests; geo; envir; Argentina; Antarc*; Antarctica
• Online Access: https://doi.org/10.1016/j.jag.2022.102862; https://doaj.org/article/4cb74b2950bf4d44add4e58cd9b6dbf4

The Araucaria-Nothofagus forests are a unique ecosystem in temperate rainforests of Chile and Argentina. They include red-listed species and have a high cultural importance for the ancestral population and thus require continuous monitoring to support conservation. Monitoring of phenology by satellite observations is a key tool to quantify the impact of climate variability on terrestrial vegetation. Here we aim to provide a first quantification and ecological understanding of the land surface phenology of protected Araucaria-Nothofagus forests in the Conguillío National Park in southern Chile. We exploit time series of enhanced vegetation index from Landsat 8 and Sentinel-2 satellite imagery from 2016 to 2020 to derive start and end-of-season (SOS and EOS) information at 10 × 10 m spatial resolution. Results show that, on average, SOS varies between 11th October and 5th November (quantiles 25% and 75% of all pixels). SOS occurs later at higher elevation, in sparsely vegetated stands, or in stands dominated by Nothofagus antarctica. EOS occurs on average between 24th March and 14th April. EOS shows a high variability between neighboring pixels that cannot be easily associated with forest stands or topography. Comparisons with regional-aggregated temperature and precipitation time series show that SOS is delayed with colder winter and spring temperatures and EOS shows stronger (but contrasting) correlations with summer and fall precipitation. By using a machine learning approach, we find that elevation is the main control on the spatial-temporal variability of SOS and EOS, followed by aspect, slope and total tree cover. These results suggest that meteorological conditions control the inter-annual variability of the phenology of Araucaria-Nothofagus forests but the effect is modified by small-scale topography, climate and stand characteristics.