Deforestación y Reforestación: Studying and Documenting the Effects of Deforestation and Subsequent Environmental Restoration Efforts in Chilean Patagonia’s Chacabuco Valley by Jackson Hotaling (25 April 2016)

Figure 1: Investigation of the recovering grasslands in the Chacabuco Valley of Chile.  Photograph by Conservacion Patagonica, 2011.



Study Area and Object of Study

Study Area

             Object of Study

Unsupervised Classification Analysis

             Gathering of Background Information

             Explanation of Process and Methodology


Learning Outcomes and Conclusions

Works Cited



Over the past 100 years across Southern Chile’s Patagonia Region, considered one of the last ‘wild places’ on earth, forest cover and natural grasslands were removed to create fenced lands for grazing by sheep and for exploitation of the lumber resources, with further harm to the environment caused by forest fires other human-related activities (  Chilean ranchers, known as gauchos, established a private ranch known as Eastancia Valle Chacabuco, which operated undisturbed for nearly a century in the Chacabuco Valley of Patagonia.  Controversy now exists in the Chacabuco Valley, where wealthy American Douglas Tompkins, founder of the North Face, the Patagonia clothing company, enters Patagonia as an outsider.  A self-proclaimed ‘eco-philanthropist,” or person who gives money or pledges to protect land in the name of environmental rehabilitation, purchased private lands from the Chilean ranchers in order to restore the natural habitat of the Chacabuco Valley in 2004.  Soon, he ordered 400 miles of fencing in the valley removed, erasing the memories of where the ranchers grew up to begin his environmental restoration efforts.  With his purchases, he initiated a process to begin creating Patagonia National Park, combining his lands in the Chacabuco Valley with the two existing National Reserves of Reserva Nacional Jeinimeni to the north and the smaller Reserva Nacional Tamango to the south.  The objective is to alter the impending loss of Coiron grasslands in the valley (to coincide with beech forest reforestation efforts in other areas of Patagonia).  According to Conservacion Patagonica, major improvements have been made—with visible advances after just five years—to return natural Coiron grasslands to the region (

This project applies remote sensing methodologies to study the effects and measure the rates of grassland restoration efforts in the Chacabuco Valley from 2004 to present.  In order to map this, Landsat-4, -7, and -8 satellite imagery are used to measure changes in the spectral resolutions of the groundcover of the Chacabuco Valley and remainder of the proposed park.  GIS software is used to identify individual pixels within the satellite images, in order to detect land cover changes over a period of several decades.  The purpose is to monitor the vegetation cover purportedly being replanted in the Chacabuco Valley, in order to assess the accurate success of the efforts by Conservacion Patagonica.  After some Chilean ranchers felt alienated by the purchase of their lands and increase in environmental tourism brought by foreign environmental organizations (who are accepting millions of dollars in private funding to reestablish the Chacabuco Valley’s beautiful landscape), this project may help discover if visible impacts have been made, and if the purchase of these lands can be properly justified by purported environmental improvements of the region.

Study Area and Object of Study

Study Area

Within the XI Aysén del General Carlos Ibáñez del Campo Region (or, simply Aysén Region) of Chilean Patagonia is the area where the proposed Patagonia National Park is currently being created.  The two National Reserves of Jeinimeni and Tamango, together 460,000 acres, combined with the former ranch Estancia Valle Chacabuco 200,000 acres large, will create a protected area of 660,000 acres, comparable to the State of Rhode Island or Yosemite National Park in California.  Estancia Valle Chacabuco is the only major landmass within the proposed park with large populaces of Coiron grasses, as it is low-lying in relation to the mountainous regions within the existing Nacional Reserves  (Figure 2).  Using ArcMap’s georeferencing tools and polygon tools, a shapefile perimeter of the proposed park was created for subsequent analysis  (Figure 3).  The park is completely within Chile, but a significant amount of the eastern boundary of the park is the Chilean border, shared with Argentina.  As a result, there is a border checkpoint within the eastern section of the park, as well as a small airstrip nearby.  Beyond camping facilities, the only major structure that will remain in existence in the park is the newly constructed Lodge at Valle Chacabuco.  At present, most of the former Estancia Valle Chacabuco ranch buildings have been removed, as well as the 400 miles of fencing that complemented the ranch and demarcated the grazing lands for sheep.



Figure 2: Extent of study area, featuring its relation to the proposed Patagonia National Park outlined in green, with an inset map featuring southern South America.



Figure 3: This map was used to derive the proposed park’s perimeter.


Object of Study

Coiron grasses are the primary grasses harmed by the constant overgrazing of sheep on the valley floor, in Latin named Festuca gracillima(Oliva, Collantes, and Humano).  These grasses initially covered the entirety of the valley and is a native species of Chilean Patagonia.  In order to protect the native grasses, and restore them to their former grandeur before overgrazing by non-native sheep, Conservacion Patagonica began work to replant vast swaths of the grassland in an effort to avoid introducing further non-native species to the region.  Environmentally aware organizations, primarily led by Conservacion Patagonica, planted these grasses across the Chacabuco Valley, and have grown impressively since the removal of the sheep and fencing.  This project aims to distinguish between the degraded grasslands and the healthy grasslands within the area of the proposed Patagonia National Park.  As a result of the overgrazing, degraded grassland, as defined in this project is generally considered any area of grasses that are not able to sustain itself.  The landscape is unnaturally rocky because healthy patches of grass were consumed and damaged by the sheep  (Figure 4).  Beginning in 2004, grasses were planted across the Chacabuco Valley and within the Proposed Patagonia National Park (Figure 5).  Healthy grassland, on the other hand, could be considered as any of these newly improved grasses, clearly flourishing in relation to the state the grasslands when the valley was still primarily used for grazing (Figure 6).



Figure 4: An example of degraded grassland, looking south from the air strip near the Chilean border post in the east of the Chacabuco Valley.  Photograph by Alexandre Rech, 2007.



Figure 5: Conservacion Patagonica exhibiting the early stages of their planting of young Coiron grasses in the proposed Patagonia National Park.  Photograph by Conservacion Patagonica.



Figure 6: An example of healthy Coiron grassland featured in a view from the central area of the valley floor.  Photograph by Hubert Zumbach, 2012.


Unsupervised Classification Analysis

Gathering of Background Information

In order to begin my analysis, a variety of sources were sought, focusing on a literature review of the involvement of foreign groups intent on improving Chile’s natural environment.  The purpose of this was to begin to understand how large of an issue eco-philanthropy was to the ranchers of the area.  Many ranchers are not pleased by the foreigners’ significant power over the lands that are owned by the eco-philanthropists, such as with the lands owned by the recently deceased Douglas Tompkins and his wife Kristine Tompkins.  A non-transparent system of land acquisition causes some gauchos to feel especially uneasy about the situation.


Explanation of Process and Methodology

 To begin the actual examination of the Chacabuco Valley itself, Landsat satellite imagery was acquired via, which makes a wide variety of satellite imagery from missions easily accessible. Landsat satellite imagery was acquired from three different missions first from Landsat-4 in1985 to develop a baseline, because the Chacabuco Valley was in private hands in 2004, and would not have faced many changes during this period.  Next, imagery was acquired from Landsat-7 in 2004, which marks the initiation of grassland restoration, and subsequently one image from every three years following to the present, in 2007, 2010, and 2013, with Landsat-8 data used in 2016.  Imagery was focused on the early summer in the Southern Hemisphere, not only for purposes of continuity, but also because images from the summer would allow for the vegetation optimum, and because rain is at its lowest levels during the summer months so reduced cloud coverage is anticipated.  With the Landsat data, one ‘True Color Composite’ image was created for compilation of images by combining the Red, Green, and Blue bands to guide with visual analysis for the experiment.

Next, the park perimeter was defined for an enhanced analysis of the images in the study period.  By georeferencing a map of the park onto the existing Landsat data, the next step involved creating a shapefile of the boundaries of the park.  This means that subsequent analysis focused on changes only within the park, rather than for the image in its entirety.  Using the park perimeter shapefile, the ‘Extract by Mask’ Spatial Analyst tool in ArcMap rid the image of all of the information outside of the park boundary.  For each Landsat band sensor in the six satellite images besides the Panchromatic band (Band 8), the data was extracted to allow for an even set of data for each year of satellite imagery analyzed.  In both the Landsat-4 and Landsat-7 imagery, there is a noticeable section of the northern part of the park that is outside of the specific image, but there are no grasslands in that area, so no other precautions need to be taken to correct that issue, as the area outside of the images will have no effect on the total percentage of grasslands within the park.

After the preliminary steps were completed for the data analysis, an unsupervised classification of each image was performed.  An unsupervised classification is a procedure completed in a GIS that groups similar pixels together within an image into clusters called classes.  Since each pixel has a unique digital number, pixels with similar digital numbers are grouped together because they share similar reflectance values across the electromagnetic spectrum, and it can be assumed that these pixels feature the same objects in them.  Each pixel within these images represents an area of 30 meters by 30 meters, so there may be several different objects within one pixel.  The unsupervised classification process simply takes the information of the object with the largest presence in the pixel in order to complete the classification.  Each ‘object’ is within its own separate class, so the groupings for the pixels will be transcribed into one of eight prominent landscape features of the Chacabuco Valley: Healthy Grasslands, Degraded Grasslands, Forest/Thick Vegetation, Exposed Bedrock, Snow/Ice, Lakes/Glacial Melt, Cloud, and No Data.  Healthy Grasslands, Degraded Grasslands, and Forest/Thick Vegetation are the object of this study, and the No Data class is composed of the aforementioned areas outside of the images being analyzed, as well as missing areas resulting from the broken Scan Line Corrector (SLC) in the four Landsat-7 images.

For the unsupervised classification, ArcMap identified 17 particular classes representing at least 100 similar pixels within each image.  These separate types of land cover, situated into classes, were manually revised and grouped into the 8 study identifiable classes found in the satellite imagery.  To discern what classes were associated with what separate types of land cover, particularly with the distinguished healthy and degraded grasslands, photographs from the study periods were used to identify in which regions experienced grassland recovery at which times.  By discovering which areas held healthy grasses, and which areas held degraded grasses, the ‘Reclassify’ tool in ArcMap could be utilized to reclassify the groups into their new classes.  The result of the reclassification is a real-world identification of separate types of land cover within the proposed Patagonia National Park  (Figure 7, Figure 8, and Figure 9).




Figure 7: Landsat-4 image, 27 January 1985.  Very small areas of healthy grasslands stood out from the Chacabuco Valley’s landscape of largely degraded grasslands.  Sheep grazing is very common during this period, harming the natural environment.



Figure 8: Landsat-7 image, 25 February 2004.  The classifications for this year did not pick up any sign of healthy grasslands, indicating that only small patches remained, but the landscape was dominated by degraded grasslands.  Also note Landsat-7’s broken SLC, which had a very small impact on grassland data in relation to the outline of the park boundaries.



Figure 9: Landsat-8 image, 9 January 2016.  Only small patches of degraded grasslands are present.  By this period, most of the 400 miles of fencing had been removed and Coiron grasses were thriving after being planted across the Chacabuco Valley.




Based on the suggestions of Conservacion Patagonica, and the evidence found in the analysis of the Landsat satellite images, environmental improvements of enhanced grasslands are visible over a period of twelve years within the proposed Patagonia National Park’s Chacabuco Valley.  These findings suggest that the purchased lands are being restored as claimed to ultimately provide the Chilean citizens (and foreign tourists) a scenic area where the natural landscape is protected.  Thus, it can also be determined that financial support for the environmental groups operating in the study region of Chile is assisting in the regrowth of the Coiron grasses in the Chacabuco Valley.

According to the unsupervised classification of the six Landsat datasets, a Healthy Grasslands improved immensely between 2007 and 2010, which corresponds to the assertion that Conservacion Patagonica identified significant improvements after five years of planting.  Using pixel count information for each completed unsupervised classification, a table and graph were created to display the changes of land cover over the study period.  Although pixel percentage information was discovered for each of the eight separate classes, only the Healthy Grasslands, Degraded Grasslands, and Forest/Thick Vegetation are presented  As exemplified on the Land Variation graph and in the analogous table, Healthy Grasslands grew to become almost 20% of the entire image, while Degraded Grasslands receded to become less than 6%, even though (Figure 10 and Figure 11).

The ecological recovery is impressive, and noticeable in photographs and by using remote sensing technology to discern substantial changes. Through a concerted effort to alter any further environmental degradation in the Chacabuco Valley, the creation of a national park could be viewed as a success  (Figure 12 and Figure 13).



Figure 10: This graph documents the changes in the study area for 3 of the 8 land cover classifications: Healthy Grasslands (with thriving Coiron grasses present), Degraded Grasslands (largely rocky landscapes that suffered from overgrazing, and Forest/Thick Vegetation (other greenery mostly existing on the lower slopes of mountains).  After planting of Coiron grasses began in the Chacabuco Valley began in 2004, noted changes occurred, with a rise of Healthy Grasslands beginning in 2010, and a fall in Degraded Grasslands after 2007.  Forest/Thick Vegetation likely did not change significantly through this period, because cloud cover existed over portions of the forested areas in 2007, and to a lesser extent in 2010 and 2013.



Table 1: Table displaying the variations in percentages of Healthy Grasslands, Degraded Grasslands, and Forest/Thick Vegetation in relation to the total area.



Figure 11: Gaucho guiding sheep in the Chacabuco Valley.  Photograph by Kaitlin McMichael, 2011.



Figure 12: Guanacos freely crossing the Chacabuco Valley.  Photograph by Conservacion Patagonica, 2011.


Learning Outcomes and Future Projects

This work addresses just one of many questions necesary to consider for full implementation of the park; further work is needed to assess the cultural and societal impacts of foreign intrusion and land ownership on the Chilean ranchers.  In conjunction with a Theory-to-Practice Grant offered by Ohio Weslayan University, four students and two faculty members will travel to Chilean Patagonia to further this, and other, research on the topic of eco-philanthropy and its many varied impacts.  The purpose of the grant will be to understand more about what conservation work is being done in Patagonia, to track existing social movements and interact with social justice-oriented groups, identify the impacts of state-controlled parklands to compare with the land use as a private ranch, and to acknowledge the concerns of the different stakeholders in the region.  By experiencing Patagonia first-hand, the grassland restoration will be visible, and more information could be gathered for future research of this region with remote sensing.



  • Arana, M. V., Gallo, L. A., Vendramin, G. G., Pastorino, M. J., Sebastiani, F., & Marchelli, P. (2010). “High Genetic Variation in Marginal Fragmented Populations at Extreme Climatic Conditions of the Patagonian Cypress Austrocedrus Chilensis.” Molecular Phylogenetics and Evolution, 54(3), 941-949.
  • Bamonte, F. P., Mancini, M. V., Sottile, G. D., Marcos, M. A., & Gogorza, C. (2014). “Vegetation Dynamics from Lago San Martín Area (Southwest Patagonia, Argentina) During the Last 6,500 Years.” Vegetation History and Archaeobotany, 24(2), 267-277.
  • Fesq-Martin, M., Friedmann, A., Peters, M., Behrmann, J., & Kilian, R. (2004). “Late-glacial and Holocene vegetation history of the Magellanic rain forest in southwestern Patagonia, Chile.” Vegetation History and Archaeobotany, 13(4), 249-255.
  • Gamboa, G. (2006). “Social Multi-Criteria Evaluation of Different Development Scenarios of the Aysén Region, Chile.” Ecological Economics, 59(1), 157-170.
  • Jones, C. (2012). “Ecophilanthropy, Neoliberal Conservation, and the Transformation of Chilean Patagonia’s Chacabuco Valley.” Oceania, 82(3), 250-263.
  • Oliva, G., Collantes, M., & Humano, G. (2005). Demography of Grazed Tussock Grass Populations in Patagonia. Rangeland Ecology & Management, 58(5), 466-473.
  • Oliva, G., Collantes, M., & Humano, G. (2013). Reproductive Effort and Seed Establishment in Grazed Tussock Grass Populations of Patagonia. Rangeland Ecology & Management, 66(2), 164-173.
  • Peace, A., Connor, L. H., & Trigger, D. (2012). “Environmentalism, Culture, Ethnography.” Oceania, 82(3), 217-227.
  • Villavicencio, N. A., Lindsey, E. L., Martin, F. M., Borrero, L. A., Moreno, P. I., Marshall, C. R., & Barnosky, A. D. (2015). “Combination of Humans, Climate, and Vegetation Change Triggered Late Quaternary Megafauna Extinction in the Última Esperanza Region, Southern Patagonia, Chile.” Ecography, 39(2), 125-140.


Select Organizations Involved in Environmental Reconstruction in Patagonia

  • Conservacion Patagonica:
  • National Forest Corporation, Chile (CONAF):
  • Parque Patagonia:
  • The Patagonia Foundation:
  • Reforestemos Patagonia: