Impacts of urban land use and land cover change on wetland dynamics in Jimma city, southwestern Ethiopia

Wetlands are commonly defined as areas that are transitional land between terrestrial and aquatic environments (Tooth & Van Der Waal 2019; Tadese et al. 2020; Mandishona & Knight 2022). They are also defined as the units of the landscape that are found on the interface between land and water bodies (Ray et al. 2022; Wang et al. 2022a; Fang et al. 2023). Wetland ecosystems are important for ecological, economic and other social values (Assefa et al. 2021; Donatti et al. 2022; Takavakoglou et al. 2022; Xu et al. 2023). Ecologically, wetland is used for water purification, soil erosion control, global climate change mitigation, recharging and discharging of ground water and habitat for biodiversity, especially as suitable ecosystems for migratory bird species (Kumari et al. 2020; Yang et al. 2022c; Li et al. 2023b; Qu et al. 2023; Rizal et al. 2023).

The local communities that live around the wetland ecosystems harvest aquatic resources such as medicinal aquatic vegetation, fishes and other importance aquatic organisms as sources of food and for income-generating purposes (Jackson et al. 2012; Dobbs et al. 2016). In addition, wetlands have been viewed as potential sites for nature-based ecotourism and leisure sites for local communities (Chan et al. 2022; Margaryan et al. 2022; Xu et al. 2022). Wetland is under pressure due to uncontrolled anthropogenic activities that deteriorate the wetland ecological balance (Angessa et al. 2019; Degife et al. 2019; Das et al. 2020). The high demand for more open space for commercial, industrial, and residential purposes driven by rapid urbanization growth significantly contribute to wetland degradation. The share of urban green spaces, vegetation and wetlands are in decline as a result of land conversion (Kadhim et al. 2022; Yang et al. 2022a; Yin et al. 2022a).

The expansion of modernized agricultural practices for the maximization of agricultural products to feed the human population has highly accelerated wetland degradation (Li et al. 2021; Ayyad et al. 2022; Wang et al. 2022a; Yang et al. 2023). For the improvement of agricultural yields, utilization of artificial fertilizers, pesticides, herbicides and modernized machine can degrade wetland ecosystems (Zou et al. 2018; Wang et al. 2022b). According to Hu et al. (2017), about 33% of wetlands globally have been lost including 4.58 million km² of non-water wetlands and 2.64 million km² of open water in 2009.

In Ethiopia, about 13,700 km² of the total land area was covered by wetland (Teferi et al. 2010; Menbere & Menbere 2018). However, wetland resources have been degraded due to inappropriate land use such as deforestation, uncontrolled grazing, mining, irrigation activities, dam and canal construction and expansion of agricultural land (Berhanu et al. 2021; Dixon et al. 2021; Zekarias et al. 2021). Several studies have been conducted to evaluate the impacts of land use/land cover (LULC) change on wetland ecosystem dynamics (Chen et al. 2022; Wang et al. 2022c; Yin et al. 2022b; Liu et al. 2023). According to Moisa et al. (2023b), due to expansion of agricultural lands, wetland area has declined at a rate of 4.2 km²/year in Abay Choman and Jimma Geneti watersheds. Consequently, wetlands and water bodies have severely declined by 1,618 ha due to expansion of built-up areas and cultivation areas in and around Bahir Dar City (Assefa et al. 2021).

Some studies have been conducted, including exploring spatio-temporal patterns of informal settlements, urban heat island dynamics in response to LULC changes, and effects of LULC changes on surface temperature in Jimma city (Moisa et al. 2022a, 2022b). However, previous studies have not considered the impacts of changes in urban LULC on wetland ecosystem degradation in relation to uncontrolled urban growth and expansion in and around Jimma city. Wider areas of wetlands have been converted to bare land, urban agriculture, and human settlement, which negatively impact the biological significance of the wetland ecosystem.

Urban expansion has also exacerbated the loss of wetland vegetation, accompanied by a significant reduction in wetland ecosystems. Beside this, other parts of the city's wetlands are used as landfills and waste disposal sites due to lack of information about and attention to the contribution of wetland ecosystems. This and other related information gaps put the wetland ecosystem under threat, which requires research attention to restore the degraded wetland and improve the wetland resource management system. Analyzing the relationship between wetland and other land cover classes such as built-up area and vegetation cover is essential for determining their negative impacts on wetland loss. Geospatial technology was used to map the spatial distribution, as the extent of wetland and its relation to other LULC classes is important for decision makers. Therefore, this study aims to fill the existing research gaps by analyzing the impacts of LULC changes on wetland degradation through correlation and regression of wetland (normalized difference water index, NDWI) with vegetation cover (normalized difference vegetation index, NDVI) and built-up area (normalized difference built-up index, NDBI) by using geospatial techniques in Jimma city, southwestern Ethiopia.

Landsat images of multispectral bands from 1992, 2000, 2009 and 2022 were used for LULC classifications. Landsat images were classified by using supervised classification with maximum likelihood algorithm (Simelane et al. 2021; Dibs et al. 2023). The LULC types of the study area were classified as: bare land, farmland, grassland, settlement (built-up area), vegetation and wetland. The detailed information on LULC types is provided in Table 2.

In the present study, the LULC change detection for Jimma city was applied by using the land use transfer matrix from 1992 to 2022. The rate change, losses and gain of LULC within the study period were calculated by using the change matrix in the study area (Pal & Ziaul 2017; Zhang et al. 2021).

Correlation and linear regression are the two techniques most commonly employed for examining the relationship between dependent (NDWI) and independent (NDVI, NDBI) variables (Miss 2020). In this study, linear regression determines the relationship as an equation, whereas correlation assesses the strength of the linear relationship between two variables (Hopkins & Ferguson 2014; Yang et al. 2022b). This analysis was used to realize the reliability impacts of NDVI and NDBI variables as major causes and effects for wetland (NDWI) degradation. ArcGIS environment and Microsoft Excel were used to analyze regression and correlation in this study.

The LULC classes of 2000 show the domination of farmland (4,611.5 ha, 43.9%) followed by grassland (3,258 ha, 31%) and wetland (1,072.2 ha, 10.2%) in Jimma city. However, vegetation (809.7 ha, 7.7%), settlement (612.4 ha, 5.5%) and bare land (146.4 ha, 1.4%) were the least dominant land cover types, respectively (Table 4). Geographically, northern and eastern parts of the study area were mostly occupied by vegetation and wetland land cover types while built-up areas highly dominated the central parts of Jimma city. However, larger parts in the southwest of the study area were dominated by extensive farmland, whereas some of the northern parts were covered by bare land (Figure 3(b)).

Farmland and grassland of LULC types in 2009 were highly dominant with 4,120.8 ha (39.2%) and 3,639.5 ha (34.6%), respectively. Besides this, vegetation and settlement occupied 1,678.8 ha (16%) and 674.9 ha (6.4%) of the total area. However, wetland was the least dominant LULC class with an area of 221 ha (2.1%) in the study period (Table 5). Except for the central and some southeastern parts of Jimma city, farmland, grassland, and vegetation covered the vast majority of the study area. Additionally, the central and southern parts of the study area were occupied with settlements and bare land (Figure 3(c)). According to Malede et al. (2023), rapid human population growth was the main cause of deforestation in the Birr river watershed of Abay basin, Ethiopia.

Land use and land cover classes of 2022 indicate that both farmland and settlement dominate wider parts of the study area with an extent of 5,075.4 ha (48.3%) and 2,625 ha (25%), respectively. Consequently, vegetation and bare land cover an area of 801.6 ha (7.6%) and 307.4 ha (2.9%), whereas wetland covers 101.8 ha (1%) of the total study area (Table 6). Geographically, settlement and farmland dominated the central, southern, some western and northeastern parts of the study area, respectively. Consequently, vegetation and wetland occupied some eastern, northern and southern parts of Jimma city, respectively (Figure 3(d)). Urban expansion has had a negative impact on water bodies and wetland degradation in and around Bahir Dar City (Assefa et al. 2021).

The overall classification exactness of the LULC precision appraisals for 1992, 2000, 2009, and 2022 was 87.5%, 89.4%, 90.7%, and 92.3%, respectively. As a result, for the think about periods 1992, 2000, 2009, and 2022, the kappa coefficients were 0.88, 0.87, 0.81 and 0.85.

The rate of LULC change was used to evaluate speed of increase or decrease in LULC types in the study area. The results show that farmland, settlement and bare land increased at a rate of 216.9, 21.3 and 8.1 ha/year and caused the reduction of vegetation, grassland and wetland cover at a rate of 101.5, 81.4 and 55 ha/year respectively from 1992 to 2000. However, farmland, settlement and bare land increased at a rate of 73.3, 70.4 and 7.5 ha/year and were responsible for a reduction in grassland, wetland and vegetation cover at the rate of 76.8, 47 and 27 ha/year, respectively from 1992 to 2022 (Table 7). The finding of this study was in agreement with the study conducted by Moisa et al. (2023b) in Abay Choman and Jimma Geneti watershed, which indicated that the extent of wetland and forest land declined at a high rate due to expansion of cultivated land. In addition, forest and grassland experienced a high rate of change due to expansion of the settlement area in Anger river sub basin from 1991 to 2020 (Moisa et al. 2022c).

Correlation analysis was used to determine positive or negative strong relationships between two variables. In this study, correlated variables were expressed as NDWI with NDVI, NDWI with NDBI, NDVI with NDBI. From the calculated value, the results showed that NDWI has a positive strong relationship with NDVI with a coefficient determination (R²) of 0.97. However, the correlation NDWI with NDBI (R² = 0.99) and NDVI with NDBI (R² = 0.99) showed a strong negative relationship (Table 9). The negative correlation between NDWI and NDBI shows that expansion of built-up area (NDBI) was the main reason for the decline of wetland (NDWI) and green vegetation cover (NDVI). A study by Guha & Govil (2022) in Raipur City, India reported that NDWI has a strong negative relationship with NDBI and strong positive relationship with NDVI.

Linear regression analysis was used to forecast the cause and effects of one dependent variable based on two or more independent variables. In this study, causes and effect of wetland (NDWI) was determined by expansion of built-up area (NDBI) and decline of vegetation cover (NDVI). The P value shows that NDBI is a better predictor of NDWI than NDVI. Our results show that substantial decline of NDWI was the primary cause of the increasing trend of NDBI. Expansion of the built-up area was another cause for the decline of vegetation cover in the study area. Calculated adjusted R² shows that expansion of built-up area and reduction of vegetation cover contributes to 98% of wetland degradation. The remaining percentage may be affected by other factors (Table 10). According to Moisa et al. (2022a), surface temperature of Jimma city increased by 96.6% due to increasing NDBI and decreasing NDVI over the study period.

Rapid human population growth was the main cause of urban expansion. This settlement may be informal or formal due to migration from rural to urban areas for changing livelihoods, which increases demand for housing in Jimma city. Anthropogenic activities, farmland, built-up area (settlement) and bare land dramatically increased. However, vegetation, grassland and wetland substantially declined over the study period. The results revealed that settlement, farmland and bare land showed increasing trends with an area of 2,111 ha, 2,199.4 ha and 225.4 ha, respectively, from 1992 to 2022, whereas grassland, wetland and vegetation showed decreasing trends with an area of 2,304.9 ha, 1,410.2 ha and 820.5 ha, respectively from 1992 to 2022.

From the calculated LULC types, farmland, settlement (built-up area) and bare land increased at a rate of 73.3 ha/year, 70.4 ha/year and 7.5 ha/year, respectively, from 1992 to 2022. However, grassland, wetland and vegetation decreased at a rate of 76.8 ha/year, 47 ha/year and 27.4 ha/year, respectively. Results show that the decline of wetland, grassland and green vegetation in Jimma city were caused by rapid expansion of built-up area and increasing urban agriculture.

From the calculated LULC change matrix, farmland, wetland and vegetation were converted to settlement (built-up area) with an area of 79.1 ha, 16.9 ha and 6.4 ha, respectively from 1992 to 2022. Bare land gained an area of 2.4 ha and 16.9 ha from wetland and grassland, respectively. However, about 2,961.8 ha was unchanged out of the total study area from 1992 to 2022 in Jimma city. The degradation of wetland in the study area was determined by NDWI, NDVI and NDBI. The parameters had strong negative and positive relationships to each other with coefficient determination (R²) of 0.98. The result indicated that wetland (NDWI) declined due to increasing built-up area (NDBI) and degradation of vegetation cover (NDVI). The linear regression analysis, cause and effect dependent variable (NDWI) and independent variables (NDBI and NDVI) were analyzed. The result shows that wetland was degraded by 98% due to the increase in built-up area and decline of vegetation cover. Based on this study, we suggest that environmental experts and natural resources managers must educate the public and promote the responsible use of natural resources, with a special emphasis on the conservation and protection of wetlands to ensure wise urban development. In addition, the impacts of precipitation and temperature on urban wetland ecosystems and the analysis of urban expansion should be studied using high-resolution satellite imagery.

The authors acknowledge Wollega University Shambu Campus, Kotebe University of Education and Jimma University College of Agriculture and Veterinary Medicine for the existing facilities to carry out this study.

YWB and DAN participated in research design, data collection, Landsat image and document analysis. TWB, BCW and GYJ participated in methodology, data analysis and interpretation. MBM, MMG and DOG participated in research design, literature review, data analysis and manuscript writing. All authors read and approved the final manuscript for publication.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.