Author(s): Ulas Qazi
Mentor(s): Edward John Oughton, Geography and Geoinformation Science
AbstractIt is estimated that more than 10 million hectares of forests are lost each year, which contribute to between 12 and 20% of total global greenhouse gas emissions. It is also theorized that a relationship exists between early economic growth of a nation and degrading environmental quality, though this is expected to balance at some tipping point. The goal of this project was to understand the relationships between deforestation, land cover change and economic growth, by seeking to determine what areas of former forests are converted to, if a correlation between deforestation and economic growth exists, and whether or not land use change can be a determinant for economic change . Analysis was conducted for three years, 2005, 2010, and 2015, using publicly available data for 20 countries selected at random to represent regional and income groups, defined by the World Bank. Initial findings suggest a correlation between rising economic prosperity and expansion of Croplands and Woody Savannas, while Barren areas dwindled when grouping data by income classes. When observing the data by region, more prosperous countries tended to see regrowth of forested areas, while less prosperous nations faced grassland expansion . To truly draw a correlation between economic growth and deforestation, and land change as an economic indicator, further analysis is needed on a global scale, though this project serves as a proof of concept to be scaled up.
Audio Transcript0:00
Hello
0:01
my name is Ulas Qazi
0:02
and this my summer research I conducted with Dr. Oughton
0:05
A spatio-temporal analysis of Deforestation
0:09
Land Cover Change and Economic Development
0:13
The key goals of the research were to:
0:15
Identify what deforested lands are converted too, and whether there is a primary type of conversion?
0:22
To Determine if there is a relationship between deforestation and economic growth.
0:27
It is often theorized that the quality of the environment degrades in early stages of economic growth/development,
0:34
though it reaches a tipping point and begins to improve
0:37
Lastly, whether or not land use change, can be an indicator for economic growth
0:44
I should mention now, that only 20 countries were studied during this period, to serve as a proof of concept,
0:49
as there is still some work to do before scaling up to a global level, but more on that later
0:55
To conduct this research, 4 sets of data were used.
0:59
Shapefiles of national and administrative boundaries,
1:02
Population
1:03
GDP per capita based on purchasing power parity
1:07
and land cover datasets.
1:10
Population and GDP data came as global coverage Tiff files
1:14
which were clipped by national boundaries, and then statistics, such as population sum
1:19
and mean gdp ppp were taken for each region within a desired nation.
1:25
The land cover dataset had a similar process though one extra step as the data came in tiles for each year covering the world.
1:32
National boundaries were compared to the bounds of rasters until an overlap was found.
1:37
Overlapping rasters were then merged and clipped to a nation, for counts of land cover type to be taken.
1:43
Results were then grouped by world bank income groups, and world bank regions in order to see if correlations existed between the variables.
1:51
Here is a description of the land cover types, feel free to pause and read over.
1:58
Results by World Bank Income Group
2:01
Every income group saw continuous growth in GDP PPP except for low income countries,
2:08
though they did have the highest percent change of any region at 43% between 2005 and 2010.
2:16
Interestingly, the extent of barren areas mimicked economic growth in low income countries, with the areas either converted to grasslands or woody savannas.
2:25
Additionally savannas, closed shrublands and evergreen broadleaf forests saw a decrease, which may suggest reliance on shrubbery and lumber (natural vegetation) for fuel.
2:35
Lower Middle Income countries saw a reduction of grasslands, though this also depended on the region.
2:41
Both Barren areas and Savanas also saw reductions while croplands rose.
2:47
Upper middle income countries had the highest rate of deforestation, as areas of evergreen broadleaf forests were lost, and woody savannas expanded
2:56
Croplands also saw expansion as grasslands reduced
2:59
This coupled with the GDP and Population increase may indicate that the countries are planning for continuous population growth, by expanding food supply, and collecting lumber for homes
3:09
High income countries had significant reductions in their two most prevalent classes Savannas and Grasslands
3:15
Savannas appear to be transformed primarily into woody savannas in Panama, while in Croatia and Portugal, the land appears to convert back and forth.
3:23
Savannas in the Netherlands continued to shrink, but further validation is needed due to the resolution error in the image, but that will be later.
3:32
Results by World Bank Regions
3:35
GDP PPP continuously grew in every region except the Middle East and North African regions,
3:42
which also contains Afghanistan and Pakistan.
3:45
Population grew in every region except Emerging and Developing Europe
3:49
Advanced economies had continuous reductions in grasslands and savannas, while woody savannas grew.
3:55
Evergreen Needleleaf forests also grew though were not a dominant class, which indicates a later stage of forest transition for the selected countries.
4:04
The Caucasus and central Asian region primarily converted grasslands to croplands.
4:09
Barren areas also reduced, and usually converted to savanas or woody savannas.
4:15
Areas with permanent snow or ice seemed to replace small patches of croplands, which may also explain the rise in croplands.
4:21
Between 2005 and 2010, the region had the highest percent change in mean GDP PPP, at 83%.
4:29
The results of Emerging and Developing Asia are slightly skewed due to only two countries falling into this region,
4:36
though we see conversion of Evergreen Forests to Woody Savannas in Malaysia and expansions of Grasslands in Bhutan.
4:42
This region is also skewed due to the water class, though that will be discussed later.
4:47
Emerging and Developing European countries had reductions in savannas and woody savannas while the extent of croplands fluctuated
4:54
There was growth in some non dominant land classes such as, Mixed Forests, Deciduous Broadleaf forests, and Evergreen Needleleaf forests which indicates that this region is past its tipping point.
5:06
Latin America and the Caribbean lost Evergreen Broadleaf Forest coverage by an average of 1.75% over each time period
5:14
Woody Savannas appear to make up the majority of converted land, though tend to border between savannas and the forests, which may be indicative of poor forest health overall.
5:24
There was a higher percent change in GDP PPP between 2005 and 2010, which coincides with greater loss of Evergreen Broadleaf forests
5:33
in the earlier period as well. Considering Honduras is included,
5:36
which is a low income developing nation, this indicates that the region has yet to reach its tipping point.
5:43
The Middle East and North African region was mostly barren, though amounts receded with each time period, and the class still covered 41.5% of the region’s total area.
5:53
Grasslands had the largest increase between each time period and appeared to most occur in areas with open shrubland.
6:00
Given the low amounts of forests and lumber in the area, open shrublands were likely stripped of small trees over time.
6:07
Grasslands and Croplands both saw growth in the sub Saharan African region, though grasslands had a greater rate over each time period.
6:15
Grasslands tended to bleed into areas of open shrubland, similarly to that in the Middle East and North African Region, indicating that lumber was once again extracted from open shrublands
6:27
Errors
6:28
As mentioned earlier, there are two errors encountered that still need solutions prior to scaling up.
6:34
Both the inclusion of water in Malaysia, and coarse resolution in the Netherlands are likely due to an issue when merging.
6:42
Islands and some nations with a coast line seem to be affected by the water issue, as a region will shift slightly.
6:49
The degradation of resolution is more random and we are still looking for a solution.
6:55
Conclusions
6:57
Based on this initial testing, Land Cover type in general, seems to have a greater influence on economic growth, rather than deforestation.
7:05
Converted cover type of deforested land is region specific naturally, though woody savannas and savannas were a common cover in all regions except for the Caucasus and Central Asian region.
7:16
However further work is required to identify regions of change within a nation and the correlation between GDP PPP.
7:24
Based on the current findings, expansion of croplands and decreasing Barren areas seem correlated to increasing GDP PPP.
7:32
Similarly, reduction of open or closed shrublands in regions where either class is a primary wooded area, may also be an indicator of economic development.
7:41
Increases in a forested land category seemed to result in slower GDP PPP growth
7:47
though those countries that had regrowth, also typically had a higher mean GDP.
7:52
In order to truly find a correlation between deforestation and economic growth, a global analysis is necessary.
7:59
While there is still work to do, we are excited about what the next few months might bring us, as the intersection of land cover changes and GDP data with a focus on deforestation is a unique problem.
8:10
Despite the inconclusive results, a lot of the groundwork has been laid, and the code created to do this processing will be released as an open source repository.
8:19
Thank you
8:27
Um, if you guys want the code right now
8:29
it’ll be right here!
Hello
0:01
my name is Ulas Qazi
0:02
and this my summer research I conducted with Dr. Oughton
0:05
A spatio-temporal analysis of Deforestation
0:09
Land Cover Change and Economic Development
0:13
The key goals of the research were to:
0:15
Identify what deforested lands are converted too, and whether there is a primary type of conversion?
0:22
To Determine if there is a relationship between deforestation and economic growth.
0:27
It is often theorized that the quality of the environment degrades in early stages of economic growth/development,
0:34
though it reaches a tipping point and begins to improve
0:37
Lastly, whether or not land use change, can be an indicator for economic growth
0:44
I should mention now, that only 20 countries were studied during this period, to serve as a proof of concept,
0:49
as there is still some work to do before scaling up to a global level, but more on that later
0:55
To conduct this research, 4 sets of data were used.
0:59
Shapefiles of national and administrative boundaries,
1:02
Population
1:03
GDP per capita based on purchasing power parity
1:07
and land cover datasets.
1:10
Population and GDP data came as global coverage Tiff files
1:14
which were clipped by national boundaries, and then statistics, such as population sum
1:19
and mean gdp ppp were taken for each region within a desired nation.
1:25
The land cover dataset had a similar process though one extra step as the data came in tiles for each year covering the world.
1:32
National boundaries were compared to the bounds of rasters until an overlap was found.
1:37
Overlapping rasters were then merged and clipped to a nation, for counts of land cover type to be taken.
1:43
Results were then grouped by world bank income groups, and world bank regions in order to see if correlations existed between the variables.
1:51
Here is a description of the land cover types, feel free to pause and read over.
1:58
Results by World Bank Income Group
2:01
Every income group saw continuous growth in GDP PPP except for low income countries,
2:08
though they did have the highest percent change of any region at 43% between 2005 and 2010.
2:16
Interestingly, the extent of barren areas mimicked economic growth in low income countries, with the areas either converted to grasslands or woody savannas.
2:25
Additionally savannas, closed shrublands and evergreen broadleaf forests saw a decrease, which may suggest reliance on shrubbery and lumber (natural vegetation) for fuel.
2:35
Lower Middle Income countries saw a reduction of grasslands, though this also depended on the region.
2:41
Both Barren areas and Savanas also saw reductions while croplands rose.
2:47
Upper middle income countries had the highest rate of deforestation, as areas of evergreen broadleaf forests were lost, and woody savannas expanded
2:56
Croplands also saw expansion as grasslands reduced
2:59
This coupled with the GDP and Population increase may indicate that the countries are planning for continuous population growth, by expanding food supply, and collecting lumber for homes
3:09
High income countries had significant reductions in their two most prevalent classes Savannas and Grasslands
3:15
Savannas appear to be transformed primarily into woody savannas in Panama, while in Croatia and Portugal, the land appears to convert back and forth.
3:23
Savannas in the Netherlands continued to shrink, but further validation is needed due to the resolution error in the image, but that will be later.
3:32
Results by World Bank Regions
3:35
GDP PPP continuously grew in every region except the Middle East and North African regions,
3:42
which also contains Afghanistan and Pakistan.
3:45
Population grew in every region except Emerging and Developing Europe
3:49
Advanced economies had continuous reductions in grasslands and savannas, while woody savannas grew.
3:55
Evergreen Needleleaf forests also grew though were not a dominant class, which indicates a later stage of forest transition for the selected countries.
4:04
The Caucasus and central Asian region primarily converted grasslands to croplands.
4:09
Barren areas also reduced, and usually converted to savanas or woody savannas.
4:15
Areas with permanent snow or ice seemed to replace small patches of croplands, which may also explain the rise in croplands.
4:21
Between 2005 and 2010, the region had the highest percent change in mean GDP PPP, at 83%.
4:29
The results of Emerging and Developing Asia are slightly skewed due to only two countries falling into this region,
4:36
though we see conversion of Evergreen Forests to Woody Savannas in Malaysia and expansions of Grasslands in Bhutan.
4:42
This region is also skewed due to the water class, though that will be discussed later.
4:47
Emerging and Developing European countries had reductions in savannas and woody savannas while the extent of croplands fluctuated
4:54
There was growth in some non dominant land classes such as, Mixed Forests, Deciduous Broadleaf forests, and Evergreen Needleleaf forests which indicates that this region is past its tipping point.
5:06
Latin America and the Caribbean lost Evergreen Broadleaf Forest coverage by an average of 1.75% over each time period
5:14
Woody Savannas appear to make up the majority of converted land, though tend to border between savannas and the forests, which may be indicative of poor forest health overall.
5:24
There was a higher percent change in GDP PPP between 2005 and 2010, which coincides with greater loss of Evergreen Broadleaf forests
5:33
in the earlier period as well. Considering Honduras is included,
5:36
which is a low income developing nation, this indicates that the region has yet to reach its tipping point.
5:43
The Middle East and North African region was mostly barren, though amounts receded with each time period, and the class still covered 41.5% of the region’s total area.
5:53
Grasslands had the largest increase between each time period and appeared to most occur in areas with open shrubland.
6:00
Given the low amounts of forests and lumber in the area, open shrublands were likely stripped of small trees over time.
6:07
Grasslands and Croplands both saw growth in the sub Saharan African region, though grasslands had a greater rate over each time period.
6:15
Grasslands tended to bleed into areas of open shrubland, similarly to that in the Middle East and North African Region, indicating that lumber was once again extracted from open shrublands
6:27
Errors
6:28
As mentioned earlier, there are two errors encountered that still need solutions prior to scaling up.
6:34
Both the inclusion of water in Malaysia, and coarse resolution in the Netherlands are likely due to an issue when merging.
6:42
Islands and some nations with a coast line seem to be affected by the water issue, as a region will shift slightly.
6:49
The degradation of resolution is more random and we are still looking for a solution.
6:55
Conclusions
6:57
Based on this initial testing, Land Cover type in general, seems to have a greater influence on economic growth, rather than deforestation.
7:05
Converted cover type of deforested land is region specific naturally, though woody savannas and savannas were a common cover in all regions except for the Caucasus and Central Asian region.
7:16
However further work is required to identify regions of change within a nation and the correlation between GDP PPP.
7:24
Based on the current findings, expansion of croplands and decreasing Barren areas seem correlated to increasing GDP PPP.
7:32
Similarly, reduction of open or closed shrublands in regions where either class is a primary wooded area, may also be an indicator of economic development.
7:41
Increases in a forested land category seemed to result in slower GDP PPP growth
7:47
though those countries that had regrowth, also typically had a higher mean GDP.
7:52
In order to truly find a correlation between deforestation and economic growth, a global analysis is necessary.
7:59
While there is still work to do, we are excited about what the next few months might bring us, as the intersection of land cover changes and GDP data with a focus on deforestation is a unique problem.
8:10
Despite the inconclusive results, a lot of the groundwork has been laid, and the code created to do this processing will be released as an open source repository.
8:19
Thank you
8:27
Um, if you guys want the code right now
8:29
it’ll be right here!
2 replies on “A Spatio-Temporal Analysis of Deforestation, Land Cover Change and Economic Development.”
Really dig the code at the end, cheers!
Interesting topic well presented. I found myself wondering if looking at types of land change might be correlated differently with GDP. Could forest regrowth be a sign of deliberate conservation measures but deforestation be driven by economic growth? Or maybe one happens early on and the other at a later stage. (I think you mentioned that in the video.) So many ways to look at the data. Thank you for sharing.