SRMP Project- Deforestation in Madagascar
Blog written by Jason Jiang and Chenjie Gu
Struggles in the Beginning
Since I did not have prior coding experience, I did not know to use R Script which is the coding software required to accomplish many of our tasks. I learned R with the Datacamp courses. I think R Script is very interesting since it can handle large amounts of data with a loop. The deforestation data was tremendously large which made it difficult to handle. If we were to include the unprotected parks, we ended up having over 150 parks!!! Each park is also multi-year data that contains 18 years from 2000 to 2018. During mid-march, our meetings were done online due to the coronavirus pandemic, this made learning to write a loop much more difficult. A loop is a large code that allows us to mask and crop all the parks at once. It was frustrating because there were multiple errors and since we’re not meeting in person, I couldn’t receive immediate help from my mentor.
Introduction
Madagascar is the world’s 2nd largest island country located in the Indian Ocean. The island harbors a range of forest types, plateaus, and deserts. It’s home to some of the world’s largest coral reef systems and the most extensive mangrove areas in the Western Indian Ocean. However, the endemic/ native species that are located on the island are sensitive to deforestation. Endemic species are species that are native only to a particular locality or region. Recent deforestation in Madagascar has been severe and is slowly accelerating. Almost 40% of Madagascar's original forested area has been lost since the 1950s. The main cause of deforestation is from social processes such as cutting wood and agricultural policies such as slash and burn which is where farmers cut down parts of the forest and then burn the area to release nutrients into the soil. Although Madagascar has over 100 protected regions, no study has examined its effect on deforestation rates. Each forest type in Madagascar has different deforestation rates.
Our objective is to find and compare deforestation rates between protected and unprotected areas in Madagascar. We want to determine whether or not protecting parks in Madagascar reduces deforestation. Hopefully, this will help encourage actions to protect Madagascar's parks.
Methods
We obtained the location and area for each protected space from the United National Environmental Programme World Conservation Monitoring Center and our remote sensing data is from the University of Maryland Department of Geographical Sciences. Remote sensing data is a high-resolution picture from satellites in space. (Landsat images from 2000-2018) The University of Maryland already processed the satellite images to classify areas of forest from non-forested areas (non-forested areas like rivers, lakes, farmland, urban areas) Remote sensing data is then changed to forest coverage and defines a region as forest if vegetation is over 5m.
We used R scripts and classified tree coverage data as humid, dry, spiny, and mangrove based on a bioclimatic envelope model and work done by a previous SRMP year. We excluded Mangrove forests because of its small area size. We then took outlines (shapefiles) of each park and overlaid it on the forest coverage map so that we can obtain the forested regions within the park. We also moved protected areas but kept it within the same forest type. We compared protected regions to similar unprotected regions via linear coordinate translation (aka overlaying forest coverage map so that we can obtain the forested regions within the park). Linear regressions were then generated to compare deforestation trends for each year for protected and unprotected areas. We also performed T-tests on log transform data to correct for skewness (specifically, most areas had very low deforestation rates, but others had very high rates, creating a skew)
Results
We compared the rates of deforestation in all 3 forest types: spiny, dry, and humid. In general, when we compared all forest types together, there was no significance between protected and unprotected. However, in humid regions, there is significance since protected regions have lower deforestation. Dry regions had no significance and spiny almost have significance. But, the most significant pattern in spiny is that protected areas have higher deforestation rates. This was not what we expected at all. Protected areas are supposed to have lower deforestation, this reveals that the protection in the parks is not working. Spiny regions are also the second-largest area and although protected areas seem to have higher deforestation, the t-tests show that there is barely any significance so we might need more data to come to a conclusion.
Discussion
We looked at 76 total parks and most of the forest area is humid. Humid regions also have low deforestation in protected parks which is a good sign. However, in spiny regions, deforestation in protected parks is higher than in unprotected areas which is a bad sign. Since dry regions do not differ significantly in terms of deforestation, we will need more data to come to a conclusion. In general, deforestation increases as time increases which shows that deforestation is getting worse. In the future, we would like to figure out why the protected areas do so well in humid regions but not in spiny regions, or do politics or management in the area lead to these differences.
Final Thoughts
From my experience at SRMP, I learned so much about science and its impact on the real world. I also learned to code from SRMP and I actually enjoyed it because of the satisfaction gained when a code you worked on works, but it can get annoying when little mistakes prevent the code from running. I would recommend SRMP to others, and not only for science lovers, but I also recommend it to anyone who has an interest in discovering more about our world and how science plays a role in it. SRMP exposes us to the knowledge that can help improve our understanding of our world/ science, public speaking skills, and also allows us to have interactions with scientists. After this experience, my interest in science as a study in college remains unchanged. In fact, I’m more inclined to study biology after my experience because I realize now how severe deforestation is in Madagascar and the importance of protecting our planet.
Blog written by Jason Jiang and Chenjie Gu
Struggles in the Beginning
Since I did not have prior coding experience, I did not know to use R Script which is the coding software required to accomplish many of our tasks. I learned R with the Datacamp courses. I think R Script is very interesting since it can handle large amounts of data with a loop. The deforestation data was tremendously large which made it difficult to handle. If we were to include the unprotected parks, we ended up having over 150 parks!!! Each park is also multi-year data that contains 18 years from 2000 to 2018. During mid-march, our meetings were done online due to the coronavirus pandemic, this made learning to write a loop much more difficult. A loop is a large code that allows us to mask and crop all the parks at once. It was frustrating because there were multiple errors and since we’re not meeting in person, I couldn’t receive immediate help from my mentor.
Introduction
Madagascar is the world’s 2nd largest island country located in the Indian Ocean. The island harbors a range of forest types, plateaus, and deserts. It’s home to some of the world’s largest coral reef systems and the most extensive mangrove areas in the Western Indian Ocean. However, the endemic/ native species that are located on the island are sensitive to deforestation. Endemic species are species that are native only to a particular locality or region. Recent deforestation in Madagascar has been severe and is slowly accelerating. Almost 40% of Madagascar's original forested area has been lost since the 1950s. The main cause of deforestation is from social processes such as cutting wood and agricultural policies such as slash and burn which is where farmers cut down parts of the forest and then burn the area to release nutrients into the soil. Although Madagascar has over 100 protected regions, no study has examined its effect on deforestation rates. Each forest type in Madagascar has different deforestation rates.
Our objective is to find and compare deforestation rates between protected and unprotected areas in Madagascar. We want to determine whether or not protecting parks in Madagascar reduces deforestation. Hopefully, this will help encourage actions to protect Madagascar's parks.
Methods
We obtained the location and area for each protected space from the United National Environmental Programme World Conservation Monitoring Center and our remote sensing data is from the University of Maryland Department of Geographical Sciences. Remote sensing data is a high-resolution picture from satellites in space. (Landsat images from 2000-2018) The University of Maryland already processed the satellite images to classify areas of forest from non-forested areas (non-forested areas like rivers, lakes, farmland, urban areas) Remote sensing data is then changed to forest coverage and defines a region as forest if vegetation is over 5m.
We used R scripts and classified tree coverage data as humid, dry, spiny, and mangrove based on a bioclimatic envelope model and work done by a previous SRMP year. We excluded Mangrove forests because of its small area size. We then took outlines (shapefiles) of each park and overlaid it on the forest coverage map so that we can obtain the forested regions within the park. We also moved protected areas but kept it within the same forest type. We compared protected regions to similar unprotected regions via linear coordinate translation (aka overlaying forest coverage map so that we can obtain the forested regions within the park). Linear regressions were then generated to compare deforestation trends for each year for protected and unprotected areas. We also performed T-tests on log transform data to correct for skewness (specifically, most areas had very low deforestation rates, but others had very high rates, creating a skew)
Results
We compared the rates of deforestation in all 3 forest types: spiny, dry, and humid. In general, when we compared all forest types together, there was no significance between protected and unprotected. However, in humid regions, there is significance since protected regions have lower deforestation. Dry regions had no significance and spiny almost have significance. But, the most significant pattern in spiny is that protected areas have higher deforestation rates. This was not what we expected at all. Protected areas are supposed to have lower deforestation, this reveals that the protection in the parks is not working. Spiny regions are also the second-largest area and although protected areas seem to have higher deforestation, the t-tests show that there is barely any significance so we might need more data to come to a conclusion.
Discussion
We looked at 76 total parks and most of the forest area is humid. Humid regions also have low deforestation in protected parks which is a good sign. However, in spiny regions, deforestation in protected parks is higher than in unprotected areas which is a bad sign. Since dry regions do not differ significantly in terms of deforestation, we will need more data to come to a conclusion. In general, deforestation increases as time increases which shows that deforestation is getting worse. In the future, we would like to figure out why the protected areas do so well in humid regions but not in spiny regions, or do politics or management in the area lead to these differences.
Final Thoughts
From my experience at SRMP, I learned so much about science and its impact on the real world. I also learned to code from SRMP and I actually enjoyed it because of the satisfaction gained when a code you worked on works, but it can get annoying when little mistakes prevent the code from running. I would recommend SRMP to others, and not only for science lovers, but I also recommend it to anyone who has an interest in discovering more about our world and how science plays a role in it. SRMP exposes us to the knowledge that can help improve our understanding of our world/ science, public speaking skills, and also allows us to have interactions with scientists. After this experience, my interest in science as a study in college remains unchanged. In fact, I’m more inclined to study biology after my experience because I realize now how severe deforestation is in Madagascar and the importance of protecting our planet.
