Additional info:

The maps are available in UTM 17S (WGS 84) projection. 
Input data for the model were two LUC Classifications from 1987 and 2001 (Meyer et al. 2011), a SRTM 90m digital elevation model (DEM), GIS-Layers of streets, villages and rivers in the study area and a GIS layer which contained the borders of the Podocarpus national park (Instituto Geográfico Militar Ecuador).
The modelled LUCC were urbanisation, deforestation and reforestation. Explanatory variables of LUCC were elevation, aspect, slope, distance from streets, distance from rivers, distance from Loja, distance from villages and distance from forest edges. The variable ?distance from forest edges? was dynamic. Therefore LUCC was not modelled directly until 2100 but with recalculation steps every 10 years. Consequently, the procedure steps 3 to 5 were accomplished for every 10 years. The Step 1 was only for some parts accomplished every 10 years.

More detailed information about the methodology used here to predict LUCC can be found in (Meyer 2010)

References:

Instituto Geográfico Militar Ecuador: Data of basically geographic featrures (rivers, roads, cities, administrative borders). Given to Rollenbeck, R. in 2002 for scientific use.)

Meyer, H. (2010): Predicting land use/land cover changes in a tropical mountain forest of Ecuador for future SVAT prediction. A modelling approach and result validation based on GIS and remotely sensed data Marburg University, bachelor thesis. 

Meyer H. , Thies B. , and Bendix J. (2011): Land use/land cover map of South Ecuador 1987 . Available online (http://www.tropicalmountainforest.org/data_pre.do?citid=1038) from DFG-FOR816dw. 

Meyer H. , Thies B. , and Bendix J. (2011): Land use/land cover map of South Ecuador 2001 . Available online (http://www.tropicalmountainforest.org/data_pre.do?citid=1039) from DFG-FOR816dw.

Additional info:

Additional info:

Additional info:

See README in data for additional information on methodology

Additional info:

Creating a classification with landsat ETM+. Preprocessing includes: 
Reprojection: Idrisi Import / Reproject, Projection: UTM 17 South (30 m pixel size)
Subsetting: Idrisi Window, min x : 684990, max x : 758010, min y: 9493010, max y: 9601610
Geometric correction: ERDAS Imagine 8.6, Geometric model: Landsat
Reference: Landsat Geocover datasets (1986, 2000, orthorectified), Digital elevation model: SRTM data, Ground Control Points: 152, Control Point Error : (X) 12.68 (Y) 13.42 (Total) 18.46 (<1 pixel), Subsetting: min x: 684990, max x: 757710,min y: 9493010, max y: 9599390
Atmospheric correction: Idrisi Atmosc (Cost model), Year: 2001, Month: 11, Day: 03, GMT: 15.00, Sun elevation:  61.8568494
Band Center (Cm) DN haze Offset Gain
1 0.485 27 -0.73807 0.11807
2 0.56 19 -0.76098 0.12098
3 0.66 12 -0.59425 0.09425
4 0.83 6 -0.60693 0.09693
5 1.65 3 -0.11912 0.01912
7 2.215 5 -0.04165 0.00665

Topographic correction: 
Statistical-empirical approach
Idrisi Hillshade
Digital elevation model: SRTM data
Sun elevation: 61.8568494
Sun azimuth: 114.5874861

CLASSIFICATION:
Step 1) Maximum likelihood classifier with 19 different training sites
Step 2) Post-classification merging to 15 and 10 classes

class  1 = Kammlage 1900-2100
class  2 = Subparamo > 2700
class  3 = Kammlagen > 2500 (gestört)
class  4 = Hanglage > 2600
class  5 = Weide
class  6 = vegetationslos (Strasse)
class  7 = ehemalige Weide (Sukzession)
class  8 = Hanglage 2200-2600
class  9 = Hanglage 1800-2100
class 10 = Kammlage 2100-2300
class 11 = Adlerfarn (Sukzession)
class 12 = Kammlage > 2300
class 13 = Weide mit Busch und Waldresten
class 14 = Rutschung
class 15 = Wasser