# ------------------------------------------------------------- #
# Datasets and Descriptions
# ------------------------------------------------------------- #

# Datasets are split into EU_west and WU_east
	EU_West: Portugal to Italy
	EU_East: Greece to Croatia


	
# ------------------------------------------------------------- #
# FOLDER: WUI Maps [WUI.rar]
# ------------------------------------------------------------- #
	# Layer
	- EU_east_WUI_2000.tif --> WUI map for 2000 for "Europe East"
	- EU_east_WUI_2020.tif --> WUI map for 2020 for "Europe East"
	- EU_west_WUI_2000.tif --> WUI map for 2000 for "Europe West"
	- EU_west_WUI_2020.tif --> WUI map for 2020 for "Europe West"
	
	# Classes
	1 = Intermix WUI (woody dominated); color code = #ED8200
	2 = Interface WUI (woody dominated); color code = #F3EA00
	3 = Intermix WUI (non-woody dominated); color code = #574935
	4 = Interface WUI (non-woody dominated); color code = #AB9322
	0 = Non-WUI
	
	# Based on Landsat-based Land Cover Fractions (Impervious, woody & non-woody fractions)
# ------------------------------------------------------------- #



# ------------------------------------------------------------- #
# FOLDER: LandCoverFractions
# ------------------------------------------------------------- #

	# Layer:
	- EU_east_2000_3BandStack.tif
	- EU_east_2020_3BandStack.tif
	- EU_west_2000_3BandStack.tif
	- EU_west_2020_3BandStack.tif
	
	# Bands: each layer represents a land cover class, scaled between 0-100%
	- Band 1: Impervious Fractions (scale 0-100%)
	- Band 2: Woody Vegetation Fractions (scale 0-100%)
	- Band 3: Non-Woody Vegetation Fractions (scale 0-100%)
	
	Visualization: Stack bands in the RGB feature space
	R: imperviousness (buildings, paved areas, ...)
	G: woody vegetation (trees, shrups, orchards, vineyards)
	B: non-woody vegetation (cropland, grassland, herbs, ...)
	
	# Raw Data Source: USGS Earth Explorer

	# Method: Spectral Unmixing using a Neural Network Regression (NNR) and syntetic Trainng data derived from Image based training data

	# Preditors: Spectral-Temporal Metrics (STM): derived from annual Landsat Time Series (Images from Jan-01 to Dec 31 of the same year). 
	The Time Series was aggregated to statistical metrics (mean, sd, Q10, Q50, Q90) which as called spectral-temporal band information and spectral indices
	
# ------------------------------------------------------------- #



# ------------------------------------------------------------- #
# FOLDER: DNDFI [DNDFI.rar]
# ------------------------------------------------------------- #

	# Layer
	- EU_east_DNDFI_2000_2020.tif
	- EU_west_DNDFI_2000_2020.tif
	
	# Scale: -1.4 to +1.4
	+1.4 --> higher positive values indicates transition towards more woody vegetation (greater distance to the 1:1 line)
	-1.4 --> higher negative values indicates transition towards more non-woody vegetation (greater distance to the 1:1 line)
	  ~0 --> values around 0 indicate low or no vegetation change and indicate being close to the 1:1line
	
	# Notes: 
	sometimes fresh burned areas can result in blue patches


# ------------------------------------------------------------- #