Text | Fisherman said
Most human activities are the main reason for the continued decline in vegetation cover on the Earth's surface, which is a factor contributing to the increase in surface temperature.
At the global level, surface temperatures are increasing due to land-use and land-cover conversions. Studies conducted in different parts of Ethiopia confirm a trend towards a significant increase in surface temperatures.
Similar to surface temperature, atmospheric temperatures increase significantly due to changes in land use and land cover.
To quantitatively describe degraded land, the Normalized Difference Poor Index (NDBaI) and the Modified Normalized Difference Moisture Index (MNDWI) are used.
Comparing LST with the Normalized Vegetation Index (NDVI), NDBaI, and MNDWI provides environmental conditions for organisms, and vegetation-based indices can indicate the presence and abundance of vegetation cover.
Spatial flexibility analysis of NDVI, LST, NDBaI and MNDWI is essential for decision-making and natural resource monitoring in natural environment surveys.
Materials and methods
1. Description of the study area
This study was conducted in the Wollega East and Horo Guduru Wollega regions between 90°27′00" and 100°18′00"N, 360°19′30" and 370°10′30"E.
Two regions, Gida Kiremu and Limu, are located in the eastern region of Wollega, while the Amuru region is administratively located in the Horo Guduru Wollega region of the Oromia National Regional State in western Ethiopia (Figure 1).
The average altitude range of the study area was 713.32 - 2496.61 m, with a total area of 5086.65 square kilometers.
2. Climate and soil type
At present, the average monthly rainfall in the study area is between 14.32 ~ 338 mm/year. The study area had more rainfall in summer, starting in June and ending in September, while spring rain was shorter, including March, April and May.
The soil types in the study area were divided into 16 types, namely: calcareous soil, calcareous dry soil, chromatic soil, chromatic soil, bitterly dissolved soil, bitterly dissolved soil, neutral dissolved soil, neutral dissolved soil, plastic dissolved soil, unisexual dry dissolved soil, fine-grained soil, northbound dissolved soil, northbound dissolved soil, northbound dissolved soil, northbound dissolved soil, northbound dissolved soil and vertical dissolved soil, among which bitterly dissolved soil dominated (2123.5 km2), and humus soil with an area of (2 km2) was the least dominant.
3. Socio-economic activities
Mixed agriculture, i.e. crop production and livestock production, is the most common source of income for smallholders. In terms of crop production, corn, peanuts, black seeds, lentils, beans and peas, as well as some vegetables (potatoes, onions, garlic) and fruits such as mangoes, papayas, oranges and bananas, are the main agricultural products in the study area.
Agricultural activities in the study area mainly rely on rain feeding.
Data type and source
Three years of Landsat imagery were used in this study. The U.S. Geological Survey provides free downloads for thermal and multispectral bands for Landsat TM 990, Landsat ETM+ 2003, and Landsat OLI/TIRS 2020.
The software used in this study is ArcGIS 10.3 and ERDAS imagine 2015.
The LST, NDVI, MNDWI, and NDBaI were calculated in this study, and the methodological flowchart of this study is shown in Figure 2.
1. Normalized Vegetation Index (NDVI)
This index is used to calculate the amount of vegetation covering the Earth's surface. NDVI is estimated using the multispectral bands of Landsat images collected during the survey.
Band 4 is used for Landsat 5 and 7 near-infrared measurements, and Band 5 is used for Landsat 8. The red band of the Landsat data was measured using band 4 of Landsat 8 and band 3 of Landsat 5 and 7. The formula for this indicator is shown in Equation 1.
2. Modified Normalized Poor Water Index (MNDWI)
When simulating thermal environments, MNDWI is designated as representative waters that typically show significant changes in thermal signature.
The formula was developed using the reflectivity values of Equation 2 in the green band (band 2 for Landsat 5 and 7, band 3 for Landsat 8) and the mid-infrared band (band 5 for Landsat 5 and 7, band 6 for Landsat 8).
3. Normalized Poor Barren Index (NDBaI)
When estimating the thermal environment, NDBaI was chosen to represent barren terrain with large differences in thermal characteristics. Equation 3 is estimated using the reflectivity of mid-infrared (band 5 of Landsat 5 and 7, band 6 of Landsat 8) and thermal infrared (band 6 of Landsat 5 and 7, band 10 and 11 of Landsat 8).
Step 1: Convert DN to brightness
Single-window technology converts digital data into AT sensor radiation sensors before calculating the luminance temperature.
The value range of TM and ETM+ DN is 0 ~ 255 (Equation (4)).
For Landsat 8, the Single Window Algorithm (MWA), used by other experts, was also employed to estimate surface temperatures. The Landsat 8 TIRS digital digits (DNs) in band 10 are first converted to spectral radiation (Equation (5)).
Step 2: Conversion to Temperature (ETM+)
Based on surface emissivity, atmospheric transmittance, brightness temperature, and surface temperature determined by single-window algorithm. The spectral radiance of TM and ETM+Band 6 imaging (as described above) can also be translated into a physically more usable quantity (Equation (6)).
The conversion formula is as follows:
In this study, the single-window algorithm (MW) was used to calculate the surface temperature of Landsat 8. It calculates the average surface emissivity and then estimates the luminance temperature from the 10-band of Landsat 8.
TB10 is a 10-band bright temperature (Kelvin K); Ɛ is the average surface emissivity (LSE) in the TIR band; W is the amount of water vapor in the atmosphere; Ɛ is the 10-band LSE of the estimated LST (Equation (7)):
Step 3: Surface emissivity estimation
According to Sobrino (2004), the emissivity is calculated using (Equation (9)).
Carlson and Ripley formula (10):
The calculated radiant surface temperature will be corrected according to the emissivity using the formula (Equation (11)):
Finally, the LST measurements for Landsat TM, ETM+, and OLI/TIRS were converted to Celsius by subtracting 273.15. (Equation (11)) converts the temperature to Celsius (°C) in units of Kelvin (K).
Results and discussion
1. Land surface temperature analysis
The spatial patterns of surface temperature in the study area were 1990, 2003 and 2020, respectively. Surface temperatures were higher in the northeast and southwest of the study area in all years (Figure 3). The increase in surface temperature in the study area was associated with a decrease in vegetation cover and an increase in bare land.
Due to the large amount of vegetation cover, the surface temperature in the eastern, central and western parts of the study area was relatively low. The average surface temperature gradually increased from 23.70°C in 1990 to 24.30°C in 2003 and 28.70°C in 2020.
From 1990 to 2020, the average temperature increased by 5°C. Moisaetal . Similar results were reported (2022a), finding that the Nu River subbasin increased surface temperature by 5.6°C between 1991 and 2020.
Compared to 1990 and 2003, 2020 was the most extreme year for temperatures. Rising surface temperatures and declining vegetation cover have led to the degradation of wetlands due to the expansion of agricultural land. Surface temperatures are expected to rise due to global warming.
2. Correlation between surface temperature and NDVI
LST values range from maximum temperature (43.2°C) to lowest temperature (16.9°C), while NDVI values range from maximum 0.50 to minimum - 0.53.
The results showed that the surface temperature was significantly negatively correlated with NDVI (R2 = 0.99). Wolteji (2022) found a moderate negative correlation between NDVI and surface temperature in the Ethiopian Rift Valley.
The results show that high surface temperature is more closely associated with low vegetation cover and vice versa.
The distribution of NDVIs during the study period is shown in Figure 4. The relationship between the two factors of LST and NDVI in 2020 is shown in Figures 2 and 5, and this study results show a strong negative correlation between surface temperature increasing with decreasing vegetation cover.
3. Correlation between LST and NDBaI
Due to the expansion of agriculture and the reduction of vegetation cover, the wasteland in the study area is increasing. The results showed that LST was significantly positively correlated with NDBaI (R2 = 0.96).
The results showed higher surface temperatures on degraded or barren terrain with higher NDBaI values (Figure 6). The relationship between the two parameters is shown in Figure 7.
4. Correlation between LST and MNDWI
Vegetation moisture content (wetlands) decreases due to the decrease of vegetation cover and the increase of surface temperature. The increase in agricultural expansion land and wasteland combined to increase surface temperature and was the main factor contributing to the decline of MNDWI in the study area.
The results showed that LST had a strong negative correlation with MNDWI, and the correlation coefficient was (R2 = 0.95). MNDWI was higher in the north and south of the study area (Figure 8).
The findings suggest that high surface temperatures are often recorded at low water levels. The relationship between LST and MNDWI is shown in Figure 9, which is consistent with the previous results.
The relationship of LST to other indices is shown in Table 2. NDBaI is positively correlated with LST, while NDVI and MNDWI are negatively correlated with LST.
5. The correlation between LST and learning time
The average surface temperature of the study area correlated with the study period. The results showed that the mean surface temperature was positively correlated with the study time, with R2 = 0.89. The average surface temperature increases over time (Figure 10).
conclusion
In this paper, we use remote sensing data to evaluate the spatiotemporal relationship between surface temperature and the NDVI, NDBaI and MNDWI indices in three regions of Jeddakirem, Limu and Amuru in western Ethiopia. From 1990 to 2020, the temperature of the Earth's surface rose by 5 degrees Celsius.
The results showed that NDVI and MNDWI were significantly negatively correlated with LST, while NDBaI were significantly positively correlated with LST. Agricultural land expansion was the main reason for the decline in vegetation cover in the study area. Declining vegetation cover is the cause of a decrease in vegetation moisture content (wetlands) and an increase in wasteland.
The surface temperature in the study area increased with time and expanded in all directions. Based on our findings, this study recommends building a sense of community to promote the wise use of natural resources to achieve their sustainability.
In addition, further research into the impact of surface temperature on agricultural production can deepen our understanding of the impact of environmental change on community livelihoods.
bibliography
[1] Aik, d.h.j., Ismail, m.h., Muharam, f.m., 2020.Land use/land cover change and its relationship with land surface temperature in Cameron Highlands, Malaysia based on Landsat and MODIS imagery, land 9,372.
[2] Bustos, E., Meza, f.j., 2015。 Methods for estimating maximum and minimum temperatures using MODIS surface temperature and vegetation data: applications in the Maipo Basin of Chile. Theorem. Acta Climatic Sinica, 36(1), 391-391.
Xiao Chen, Hui Zhao, Peng Li, Zhong Yin, 2006.Relationship between urban heat island and land use/cover change based on remote sensing imagery. Journal of Remote Sensing and Environment,34(4),133-146.]
[4] Xiao Cheng, Bin Wei, Gang Chen, Chao Song, 2015.The influence of park scale and its surrounding urban landscape pattern on the cooling effect of parksJ.Urban Planning.Dev. 141, A4014002.