Impact of tree canopy on thermal and radiative microclimates in a mixed temperate forest: A new statistical method to analyse hourly temporal dynamics

Forest shelter buffers microclimate, decreasing daily ranges of solar radiation and temperature, yielding higher minimum and lower maximum temperatures than those of open field. The most common way to analyse sets of these data is to compare mean, maximum and minimum values of climate parameters of open field and understorey conditions at daily, monthly or seasonal scales; however, this approach loses information about temporal dynamics. This study developed a statistical method to analyse hourly dynamics of temperature (T) and radiation (Rad) together and quantify effects of canopy openness and seasonality on these dynamics. Eight experimental sites were chosen in small gaps located in a temperate oak-pine forest (France), and five plots were established in each along a light gradient (i.e. a total of 40 plots), which delimited a transect along which T and Rad were measured hourly at a height of 200 cm from May 2009 to March 2010. T and Rad were also measured in open field. A specific Principal Component Analysis (PCA) with an innovative graphical representation was performed on this dataset. This statistical method allowed hourly temporal dynamics of all data recorded to be analysed and included a chart to interpret the distribution of the data in the principal plane defined by the PCA. Except in winter, results demonstrate the well -documented buffering effect of the tree canopy on T, with higher minimum and lower maximum values in the forest understorey. This effect was especially pronounced for minimum T and increased as canopy grew denser. In summer, T remained higher than expected in the understorey and was lower than expected in the open field, indicating thermal inertia in the understorey and an a priori cooling effect linked to wind or radiative losses during the night in the open field. The newly developed statistical method offers an innovative approach to better understand the tree canopy's buffering effect on temporal dynamics. (C) 2017 Elsevier B.V. All rights reserved.