GIS and Remote Sensing: Mapping Land Cover Changes in Mato Grosso State from Landsat Satellite Imagery

Posted: August 26th, 2021

GIS and Remote Sensing: Mapping Land Cover Changes in Mato Grosso State from Landsat Satellite Imagery

Name

Institutional Affiliation

GIS and Remote Sensing: Mapping Land Cover Changes in Mato Grosso State from Landsat Satellite Imagery

1.      Introduction

Deforestation causes significant changes in the land cover. The situation is made worse by significant carbon emissions. However, the increasing anthropogenic activities in tropical belts are becoming adverse as there are major forest degradation activities further contributing to carbon emissions. For instance, the Brazilian Amazon currently experiences over 40% of carbon emissions because of deforestation (Lillesand, Kiefer, & Chipman, 2013). Monitoring forestland cover is critical since increased deforestation contributes to a reduction in carbon stocks in tropical forests. According to Mass et al. (2006), selective logging is estimated to produce carbon emissions amounting to 12% of the aggregate forest degradation. Like in the case of the Amazon region, selective logging and occasional fires induce degradation of forests. It is hence critical to monitor land cover changes to ensure that sufficient strategies are defined to help curb any adverse impacts such as landslides and other epidemic casesthat are dangerous to the current and future environment as well as the population (Lillesand, Kiefer, & Chipman, 2013). Without forests, the world would continue experiencing unfavorable and drastic climatic changes. Brazil is critical in this case because it holds the largest rain forest in the world, whose impact is widespread across the world.The study conducts a spatial resolution of satellite imagesto map the changes in land cover in Mato Grosso State, Brazil,focusing on the transformation in landscape features over time.Geographical Information System (GIS) and remote sensing tool are used to conduct mapping as it can adequately cover dynamic changes in the environment. The monitoring is based on three years, that is, 1990, 2005, and 2015 to see the changes in the land cover in these periods.

2.      Methods and Study Area

The study was conducted in the Mato Grosso State area, Brazil. The area has a total land size of about 903, 365 square kilometers ( (Mass et al., 2006). The area is characterized by variable climatic conditions, terrain relief and precipitation patterns thus conducive for different sets of vegetation. Further, based on its location in the heart of deforestations within the southern parts of Brazilian Legal Amazon, the area experiences the highest rates of deforestation annually as recorded since the 1980s (Mass et al., 2006). There is also intense activities that largely contribute to forest degradation, such as selective logging and frequent forest fires. Subsequently, these activities have led to habitat fragmentation making, thus increasing the vulnerability of the forest to degradation agents. For instance, the border areas of the forest cover have a high chance of experiencing forest fires. Within Mato Grosso, fires are majorly experienced during the dry season in June and October, which accounts for an average of 85% of the overall deforestation activities in a year (Congalton & Green, 2009). In this study, all the available Landsat TM images that are available in the cloud-free.  Figure 1 below is a display of the study area

Figure 1: Study Area – Mato Grosso State with Location of Sample Sites

The method involves acquiring of Landsat images from selected systematic sample sites. The steps undertaken to capture the changes in land condition include pre-processing of satellite images, production of the fractional images of the vegetation and shades as well as creation of spectrally and spatially homogeneous areas (Schowengerdt, 1997). The forest degradation due to fires was estimated by relying on the data sets acquired from the seventy-five (75) sample sites of the survey. It is based on a systematic distribution at a longitude-latitude confluence point in the Mato Grosso area. The images in the sample sites are indicated as a burned area or selective logging areas. The information collected on these cases is used to statistically analyze the overall burned size of the forest and the affected areas due to logging to show the extent of degradation over time.

3.      Results

The burned areas observed during mapping were mostly related to either degradation or deforestation process. During deforestation, the forest is usually cleared through logging. The remaining vegetation is then burned to convert the remaining land into agricultural use (Congalton & Green, 2009). While during degradation, uncontrolled fires are used to clear the forest cover without logging. Besides, there is no conversion of the land into any other form of use (Congalton & Green, 2009). Hence, the remote sensing tool becomes essential in this case to help differentiate between forest degradation and deforestation processes. On the tool, the areas that are deforested appear as non-forest (grassland or cropland) in the successive years while the degraded forest (burned area) usually recovers as the forest grows again.

Selective logging is interpreted following the detected reduction of canopy cover and logging tracks due to logging activities. These features are also highlighted by soil fraction to facilitate the visual interpretation. The following figures, 2, 3, 4, 5, and 6, show the visual interpretation process for the three maps, displaying land changes overtime.

Figure 2: Thematic Map for 1990

The map displays the state of the forestland based on the size of the cover of different vegetation during 1990. As observed, the northeastern parts are densely forested, but as one moves down to southern parts, forest density reduces due to widespread clearance, cases of bare land, and shrubby land covers characterizing the high level of deforestation and degradation in the period.

Figure 3: Thematic Map for 2005

The map displays the state of the forestland based on the size of the cover of different vegetation during 2005. As observed, the forest density is almost of the same size as the bare land implying that the state of deforestation and degradation has taken effect by then.

Figure 4: Highlight Changes between 1990 and 2015

Figure 4 summarizes the observation made between 1990 and 2015. Overtime, degradation and excessive deforestation left much of the northern parts of the forestland highly affected with scarce vegetation. However, some of the southern parts are slowly recovering as exhibited in gradual increase in forestland.

Figure 5: Highlight Changes between 1990 and 2005

Figure 5 summarizes the observation made between 1990 and 2005. From the map, it is indicated that there are extreme levels of degradation and deforestation that occurred during this period, which have caused almost the whole forest area loss its original vegetation. Forestland has largely decreased with few parts in the southern section remaining unchanged.

Figure 6: Highlight Changes between 2005 and 2015

The map shows the state of forestland changes between 2005 and 2015. As shown in the map, the forest appears to have regrown as indicated by an increased level of vegetation. There is an increase in areas of forest cover, implying reduced deforestation and degradation. Over the period, the distribution of land coverfor Mato Grasso State area is as shown under Table 1 and 2

Table 1: Distribution of Land Cover in Mato Grasso State

	1000	% Cover
Tree Cover	439.1	49.1
Mosaic Tree Covers	24.9	2.8
Other wooded lands	90.1	10.2
Water	4.1	0.5
Other land covers	348.1	39.5

Table 2: Distribution of Burned area in Mato Grasso

	1000
Tree Cover	16.3
Mosaic Tree Covers	2.8
Other wooded lands	6.2
Other land covers	59.2

•  Discussion
  • Method application

            Non-forest and forest covers areessential when developing procedures for mapping and detecting degradation of a forest. With the help of a mask, the changes that are detected within the sample sites could easily be contextualized within the forest disturbance (Congalton & Green, 2009). While using the Landsat TM/ETM + datasets, it is easy to detect cleared and burned areas (Schowengerdt, 1997). The researcher used the remote sensing tool to examine and distinguish various features exhibited in the dynamic changes of the land cover over time (Congalton & Green, 2009). As such, these areas are mapped using semi-automated methods. Besides, it is critical to understand that the degradation of the forest because of fires, that is, burned forest is effectively mapped by using time-series images. The reason is that sometimes fire is used to clear part of vegetation, which makes it easy for one to mix the cases if the appropriate tool is not applied.

Additionally, detecting and mapping selective logging can be effectively executed by relying on soil fractioning of images (Congalton & Green, 2009). The process is effect via medium resolution. Thus, these processes were performed to ensure that quality images are attained (as revealed under the figures above) to ensure that there are sufficient details for study. The proposed method for undertaking the study was effective in facilitating the mapping of logged areas from the burned one, which allows estimating carbon emissions across a set of forest covers.

• Mapping of Mato Grosso Area Forestland Cover

Deforestation and degradation have been prevalent over the observed period. According to the 1990 mapping, the report indicated much of the northern part of the area hold high forest density (Mass et al., 2006). However, the southern part appears to be slowly degrading. Subsequently, according to the highlighted changes between 1990 and 2005, the area appears to have drastically changed as exhibited by an increase in forestland cover (Congalton & Green, 2009). These changes are attributed to the increase in agricultural activities that have been increasing overtime. The changes in the Mato Grosso area presents an example of human-induced deforestation for agricultural purposes. Since 1990, Brazil’s agricultural production has been high on the world market in products such as barley, wheat, and corn, among other staple food (Mass et al., 2006). Hence, the activities have come at the expense of losing the forestland in the Amazon, which was being cleared to expand the agricultural field.

Conversely, the 1990 and 2015 highlight depicts an entirely different picture. Most of the area appears to have decreased in terms of vegetation. The situation is attributed to cases of burned forest that yields to degradation of the forestland. In some cases, the degraded land has turned shrubby, implying evident loss of value due to uncontrolled wildfires. However, the 2005 and 2015 displays attempt to recover the forest primarily due to increasing density in the northeastern parts of the area. The changes are attributed to the introduction of government policies that reduced deforestation (Mass et al., 2006). Besides, a moratorium on exporting soy grown on deforested land served as a significant strategic intervention, thus curbing on high cases of deforestation in the area.

• Conclusion

            The topic of study was an assessment of the Mato Grasso area, which focused on the situation of land cover changes over time. The period of study was split into three sections, the first one looking at 1990, the second highlighting the changes in land cover between 1990 and 2005, the third one assessing land cover changes between 2005 and 2015, and lastly, the same issue was assessed for 1990 and 2015. The study relied on GIS and Remote sensing tools to help map out transformations in the landscape features, distinguishing between degraded and deforested landscape and its effect on the area of study. The tool applied was useful in analyzing the case, ultimately revealing enough data for interpretation. Besides, the stud established that there had been changes in land cover over the period, highlighting major cases of human-induced deforestation to expand agricultural activities. Equally, degradation was distinct, which has caused wastage of primary carbon sources and reducing the land cover to barren fields. However, government interventions appear to help the area regain its forestland, thus giving opportunity for original vegetation to regrow.

References

Congalton, R. & Green, K. (2009). Assessing the accuracy of remotely sensed data: principles and practices. Boca Raton: CRC Press/Taylor & Francis.

Lillesand, T., Kiefer, R. & Chipman, J. (2013). Remote sensing and image interpretation. Hoboken, NJ: John Wiley & Sons.

Mass, D. C., DeFries, R.S., Shimabukuro, Y.E, Anderson, L.O., Arai, E., de Bon Espirito-Santo, F., & Morsette, J. (2006). Cropland expansion changes deforestation dynamics in the southern Brazilian. International Journal of remote sensing, 20 (1), 139-152.

Schowengerdt, R. (1997). Remote sensing, models, and methods for image processing. San Diego: Academic Press.