Assessing the impact of Land use changes on ecosystem services in the Alledighe rangeland, Ethiopia

Abstract

Understanding the ecological, social, and economic values of protected areas, as well as assessing the services they provide to both humans and the environment is crucial for informing conservation policies and sustainable land management practices. Using the benefits transfer method, changes in ecosystem service values (ESVs) resulting from spatiotemporal land use dynamics were evaluated in the Alledighe Wildlife Reserve (AWR) spanning from 1998 to 2016. Five distinct habitat types, namely grassland, bushland, woodland, riverine forest, and highland forest, were identified across the landscape. The ESVs were estimated using regional and global ESV values. A decline in the extent of grassland, woodland, and riverine forests by 9.9%, 2.4%, and 1.5%, respectively, was observed while bushland and highland forests increased by 10.6% and 3.3%, respectively. The AWR experienced a loss of roughly 145 km² of grassland habitat. Based on regional and global ESVs, total ESVs in the study area decreased by 28.18% from approximately US$ 180 million to approximately US$ 129 million, and by 40.85% from approximately US$ 496 million to approximately US$ 293 million. As per individual ESV assessment, the total ESV decreased by 41% from around US$ 374.5 million to US$ 264.8 million. Provisioning service declined by 41.6% from US$ 100 million to US$ 70.6 million. Regulating service declined by 42.5% from US$ 242.4 million to US$ 170 million. Supporting service declined by 67% from US$ 5.3 million to US$ 3.2 million, and cultural service decreased by 27.8% from US$ 26.7 million to US$ 20.8 million. The larger ESV change was contributed by the expansion of forestland and bushland across previously grassland-dominated areas. The results of this study could render the value of the rangeland more visible in the decision-making process, as well as provide valuable input for future planning and management interventions of the AWR's pristine rangeland, thereby enhancing ecosystem services and the livelihoods of the surrounding communities.

1. Introduction

Spatiotemporal Land Use Land Cover (LULC) changes can be the primary factor of biodiversity loss [1,2] as well as a potential trigger for natural disasters like flooding and drought [3] and the most important factor influencing global and local environmental changes [4,5,6,7,8]. LULC changes could adversely impact biodiversity and ecosystem services [9] and ecosystem processes [10]. Further, influences vegetation distribution, soil, and water availability [[11], [12]]. Land use changes initiated by external and internal drivers have had an impact on traditional rangeland management in pastoral areas [13,14].

Changes in LULC have been increasing at global and local scales [15] and influencing biophysical systems at all levels [16]. It has a greater impact on natural ecosystems and increases human vulnerability to climate change and socioeconomic crises [16,15,17]; LULC change can be driven by a variety of environmental and social factors [18]. The major factors of LULC changes in Ethiopia are multifaceted; however, contributing factors to ongoing rapid LULC change include a lack of effective land use policy [19] and population growth [20]. Rapid economic, social, and environmental transformations in Ethiopia have increased the demand for socioeconomic necessities, putting pressure on natural areas[[21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31]].

Likewise, most of Ethiopia's protected area systems, including the Bale Mountain Eco-Region [32], Awash National Park [33], Simien Mountains National Park [34], Kafta Shiraro National Park [35], Babile Elephant Sanctuary [36], Alledighe Wildlife Reserve [37] and Nech Sar National Park [38] have considerably changed in terms of LULC. According to these studies, a large portion of grassland habitat has been encroached upon and converted to bushland. These protected area systems face several conservation challenges [39]. The continuation of degradation, primarily highly associated with farmland expansion and intensive grazing land in and around protected areas has hindered the protected area's effectiveness [40,39]. The change in LULC types combined with poor range management practices resulted in rangeland degradation in the Afar region [41]. Herding livestock searching for pasture and water is the prime income for Afar pastoralists [42]. The AWR was created to serve as a wildlife refuge between Awash National Park and Awash West Wildlife Reserve in the 1960s [43,44] and to conserve and propagate wildlife [45]. The reserve was chosen to serve as a corridor between the park and the surrounding plains, thereby protecting Awash National Park's wildlife population through migration [44]. It also serves as a charismatic wildlife area and a reliable rangeland for livestock [42,46].

There has recently been a surge in research on ecosystem services assessment, which includes ESVs for conservation planning [[47], [48]], environmental accounting [49], and land management [50]. ESV and cost estimates of ecosystem degradation are critical for incorporating the value of ecosystem services into policy and planning [51], but these are difficult tasks [52]. Most importantly, the data can be used to underpin the socio-ecological patterns of ecosystems and to forecast the effects of land use management on ecosystem services [53]. Assigning ESVs to different LULC types has been widely used as a proxy to quantify and map ecosystem services [52,54,55,56,57,58,59].

In many natural areas, including protected areas, the value of environmental services exceeds the value of direct benefits from tourism and employment [60]. A lack of understanding of the value of protected landscapes can lead to the ineffective promotion of an ecosystem [61]. The availability of reliable information on the values of ecosystem services can increase the visibility of PAs at the highest levels of decision-making [53,62,63,64,65]; increase public awareness [66] and imply land use management schemes [67,63]. Therefore, this assessment could assist protected area authorities in integrating biodiversity and ecosystem services in planning and demonstrating the landscape benefits to local communities and the national economy [68]. Further, conservationists frequently face difficulties obtaining financial and social support for the protection of landscapes that do not demonstrate societal benefits [69,70]. The ecosystem services approach helps to conserve biodiversity [[71], [72], [73], [74]] by providing sound justification as well as increased and diversified conservation funding [69,75]. As a result, this study aimed to analyze ESV changes in response to spatiotemporal land use dynamics in an iconic protected area in northeast Ethiopia, to use the information to integrate biodiversity and ecosystem services into the national agenda and imply the need for an ecosystem services approach to prioritize protected areas and accelerate the globally agreed biodiversity framework.

2. Materials and methods

2.1. The study area

The AWR is located in the Great Rift Valley along the northeastern region of the country between 39°30′ to 40°30′ E and 8°30′ to 9°30′ N. The altitude ranges between 776 m and 2445 m a.s.l (Fig. 1). The area is characterized by a semi-arid ecosystem with annual rainfall ranging between 400 and 700 mm [76]. The mean annual rainfall is 493 mm; the least occurs between October and December, and the maximum during July and August [77]. The mean seasonal temperature ranges from 25 to 30⁰ C but the daily maximum temperature may be as high as 38⁰ in June while the minimum daily temperature can drop to 15⁰ in December [77].

Fig. 1.

Location map of the Alledighe Wildlife Reserve.

The major soil types in the AWR are Vertic Cambisols and Calcic Cambisol. The chemical analysis showed that the pH of the soil is alkaline and ranges from 8.1 to 9.1. The organic matter (OM) content is low, ranging from 0.5 to 1.9% [78]. The vegetation types include grasslands, bushland, wooded grassland, riverine forest, and highland forest [79,76,42]. About 268 plant species and two types of ecosystems Dry evergreen montane forest and Acacia comiphora ecosystems [79,76], 31 mammals [80,43,81,77], and over 140 bird species have been recorded [80,43,81,77]. The Grevy zebra (Equus grevyi), Beisa oryx (Oryx beisa beisa), Soemmering's gazelle (Gazella soemmering), Gerenuk (Litocranius walleri), and Lesser kudu (Tragelaphus imberbis) are among the wild animals inhabiting the reserve [42,80,43].

Historically, the grassland plain stretching from the center of the reserve to the northwest was mainly occupied by grasses and occasionally other herbs [42,82]. The bushland is an extensively increasing habitat type that possesses an assemblage of trees and shrubs. The woodland habitat is mainly occurring in the southern, eastern, and northern edges of the landscape [42,82]. The eastern mountainous section of the landscape is characterized by dense highland forest [42]. The riverine forests are a unique ecosystem and are important for the wild animals of the landscape [79].

2.2. Methods

2.2.1. Data collection

Landsat 7 and Landsat 8's Operational Land Imager (OLI) and Thermal Infrared Sensor (TIRS) images, boasting a resolution of 30 m, were employed in the detection of LULC alterations between the years 1998 and 2016. These years were selected for analysis due to the desire to utilize as much quality data as possible. Further, going as far back as possible to monitor trends of wildlife management policy transformation [63,66] and changes that occurred following the transfer of responsibility for managing the protected area to the federal Ethiopian Wildlife Conservation Authority in 2008 can also be attributed to this study.

2.2.2. LULC classification

Landsat images obtained from the US Geological Survey (USGS) website were acquired as compressed files for the years 1998 and 2016 and were subsequently converted to TIFF format. The data underwent projection to the Universal Transverse Mercator (UTM) projection system zone 37 N with a datum of World Geodetic System 84 (WGS84) to guarantee consistency and compatibility [63].

The entire suite of images underwent classification using a supervised classification method with a maximum likelihood classifier [63], whereby representative ground truth data were gathered to train classifiers and associate the reflectance property with the object, thus enabling accurate automatic classification [83,84].

2.2.3. Image processing

The reference data were compared to evaluate and quantify the accuracy of a classified image [85]. By comparing the classified map to reference data, this method was used to estimate the accuracy of image classification, and a field study was also conducted with key informants from the local community to confirm the accuracy of the land classification undertaking. As a result, an accuracy assessment including overall accuracy, user accuracy, and producer accuracy was conducted [63] (Fig. 2; Table 1).

Fig. 2.

The methodological flow chart of the study.

Table 1.

Accuracy assessment.

Habitat types	1998		2016
	Producer	User	Producer	User
Grassland	75	95.12	75.6	87.5
Bushland	66.6	59.45	86.2	73.5
Highland Forest	63.6	82.35	100	87.5
Riverine forest	50	61.53	100	100
Woodland	85	61.81	70.5	75
Overall accuracy	71.77	71.77	81.14	81.14
Kappa statistics	0.63	0.63	0.75	0.75

2.2.4. LULC change detection

The comparison of LULC change during the study periods was executed using Equation 1 as described by Ref. [86]. Additionally, Equation (2) was employed to compare the rate of change in each LULC type within the study period. The identification of the directions of changes in each land cover type was determined by subtracting the final year from the initial year of each land cover type, then dividing the result by the initial year of each land cover type, and finally multiplying by 100, as described by Refs. [38,54].

Total area change = Area _{final year} – Area _{initial year} 1

Where an area is the extent of each LULC type; positive values suggest an increase whereas negative values indicate a decrease.

$$\text{Rate}\text{of}\text{change}=\frac{{\text{Area}}_{\text{final}\text{year}-}{\text{Area}}_{\text{initial}\text{year}}}{N}$$ (Eq. 2)

Where N is the time interval between the initial and final years.

2.3. ESVs of LULC types

For this study, the global, regional, and Economics of Ecosystems and Biodiversity (TEEB) ecosystem service values (ESVs) assigned values for distinct LULC types were utilized. The former was based on [52] ESVs, which used a representative biome as a proxy for each LULC type. Later ESVs were more conservative, with expert knowledge of the study landscape conditions and other studies, primarily from the TEEB valuation database, being used to make changes [57,58].

ESVs for the representative biome were included in the TEEB valuation data [87] (Table 2). Despite an inconsistency between LULC type and the proposed biome by Ref. [88], data obtained from Ref. [58] on individual ESVs were incorporated into the East Africa individual coefficient [87] to estimate the individual ESV [63] (Table 3).

Table 2.

Reginal and global estimates of ESVs in million US$ for each LULC type.

LULC types	Equivalent biome
		Regional	Global
Bush-shrub-woodland	Tropical forest	987	987
Grassland	Grass/rangeland	293	4166
Natural forest	Tropical forest	987	5382

Table 3.

Ecosystem functions value coefficients of each LULC type database TEEB of McVittie and Hussain (2013).

Ecosystem function types	ESVs (US$ ha−1year−1 in each LULC)
	Forest	Grassland	Woodland	Bushland/Shrubland
Water supply	178	114.4	4.33	114.4
Food production	67	130	0	130
Flaw material	475.51	19	138.2	19
Biomass fuel	31.34	79	67.28	67.28
Genetic resources	327.68	0.01	0	0.01
Water regulation	15.6	5	0	5
Water purification	275.2	99.71	253.56	99.71
Erosion control	360	55.89	26.76
Biological control	20.62	30	0	30
Biodiversity protection	263	52.68	11.44	52.68
Climate regulation	569	121.91	383.72	121.91
Gas control	22.12	2.41	0	2.41
Carbon sequestration	1229.79	297.39	8.16	297.39
Nutrient recycling	11.61	0	0	0
Pollination	48.29	32	0	0
Soil formation	3.26	7	0	0
Habitat/regulation	28.057	0	0	0
Recreation	810.68	3.11	0	0
Tourism	73.74	0	0	0

2.4. ESV analysis

The analysis of ESVs was conducted on the assigned LULC types, wherein the area of each type was multiplied by its corresponding value to obtain ESV estimates for distinct LULC types. This methodology was adopted as outlined by Ref. [88] and is represented by the formula in Eq. (3).

ESV = ∑ (A_k * VC_k) 3

Where ESV = Total estimated ES value, A_k = Area (ha) and VC_k = Value ESV (US$ha⁻¹year⁻¹) for LULC type ‘k’,

To investigate the variations in value over time, the estimated ESV was adjusted to the year 2022, using the CPI Inflation Calculator, as suggested by Ref. [63]. Additionally, the ESV change between 1998 and 2016 was determined using Eq. (4) [89,56,65].

$$ESVchange(\%)=\left(\left({\text{ESV}}_{\text{final}\text{year}}-{\text{ESV}}_{\text{initial}\text{year}}\right)/{\text{ESV}}_{\text{initial}\text{year}}\right)*100$$ 4

Where ESV = Total estimated ES value. Where positive values suggest an increase whereas negative values imply a decrease in amount.

2.5. Elasticity of the ESV change with LULC

Given that value, coefficients cannot be used to parametrize accurate ESVs, and biomes may not always be exact matches for LULC types, the elasticity of ESV analyses was conducted to determine the percentage change in ESVs for a given percentage change in the value coefficient [90]. Specifically, the estimated ESV of a particular LULC class was modified by 50% [91], while the remaining ESVs identified for LULC types were held constant. To calculate the Coefficient of Sensitivity (CS) value, a simplified method was utilized [92]. It is important to note that the ESV changes of the CS value should be less than one (Eq. (5)).

$$\text{CS}=\frac{{\text{VC}}_{\text{ik}}{\mathrm{A}}_{\mathrm{k}}}{{\text{ESV}}_{\mathrm{i}}}$$ 5

Where ESVi is the total ESV (US$ yr⁻¹) from all LULC types identified at i year (US$ yr⁻¹), VCi, k is the total value of ES provided by the k LULC class at i year (US$ yr⁻¹), and Ak is the area coverage (ha) of the k LULC type at i year.

3. Results

3.1. LULC changes (1998–2016)

Each of the five major LULC types is represented by its own set of results: grasslands, bushland, woodland, riverine forest, and highland forest. The study found that important changes in land cover occurred in the AWR between 1998 and 2016. Grassland, woodland, and riverine forest cover have decreased by 9.9%, 2.4%, and 1.5%, respectively, while bushland and highland forest cover have increased by 10.6% and 3.3%, respectively (Fig. 3; Table 4).

Fig. 3.

LULC of the AWR in 1998 and 2016.

Table 4.

LULC changes in the AWR (1998, 2016).

Habitat type	1998		2016		Change		Trend
	Area (km2)	%	Area (km2)	%	Area (km2)	%
Forest	96.5	6.6	144	9.9	47.5	3.3	Increasing
Woodland	593	40.6	558	38.2	−35	−2.4	Decreasing
Grassland	359	24.6	214	14.7	−145	−9.9	Decreasing
Riverine forest	63.5	4.4	41	2.8	−22.5	−1.5	Decreasing
Bushland	347	23.8	502	34.4	155	10.6	Increasing
Total	1459	100	1459	100

The land transition matrix presented in the AWR during the study period revealed the transformation of a substantial amount of woodland into grassland, highland forest, and bushland. The riverine forest experienced a considerable conversion into woodland and bushland. The grassland has undergone a substantial conversion into bushland. A considerable portion of the highland forest has been transformed into woodland and bushland. Lastly, the bushland has predominantly been converted into woodland (Table 5).

Table 5.

Transition matrix of LUCC in the AWR between 1998 and 2016 (unit: km²).

Land class 2016							Total
Land class 1998	Land class	Woodland	Riverine forest	Grassland	Highland forest	Bushland
	Woodland	309.55	6.44	70.66	81.09	124.04	591.79
	Riverine forest	25.00	17.69	0.95	9.25	10.61	63.49
	Grassland	31.47	11.44	131.87	1.88	182.01	358.68
	Highland forest	47.57	0.78	0.03	35.63	14.05	98.06
	Bushland	144.21	5.00	10.81	17.14	169.27	346.42
	Total	557.8	41.4	214.3	145.0	500.0	1458.4

3.2. ESVs dynamics in response to land use changes (1998, 2016)

In the study area, total ESVs decreased. According to the regional ESV, the value decreased by 28.18% from approximately US$ 180 million in 1998 to approximately US$ 129 million in 2016 (Table 6). The global ESV estimate also revealed a 40.85% drop in total ESVs from approximately US$ 496 in 1998 to approximately US$ 293 million in 2016 (Table 6). According to the regional estimate, woodland was the dominant proportional contributor to ESVs in 1998, accounting for approximately US$ 88.3 million (49.15%), followed by bushland, which contributed approximately US$ 51.7 million (28.76%) (Fig. 5). However, grassland, forest, and riverine forests make up only 8.83%, 8%, and 5.26% of the total.

Table 6.

Regional and global ESVs in million US$ for each LULC type and their proportion (1998, 2016). (The value is based on 2022 US$).

Habitat type	Regional ESV				Global ESV
	1998 US$/yr	%	2016 US$/yr	%	1998 US$/yr	%	2016 US$/yr	%
Forest	14,382,070.50	8.00	14,212,800.00	11.00	78,423,813.00	15.81	77,500,800.00	26.42
Woodland	88,378,941.00	49.15	55,074,600.00	42.64	88,378,941.00	17.82	55,074,600.00	18.77
Grassland	15,883,237.00	8.83	6,270,200.00	4.85	225,834,694.00	45.53	89,152,400.00	30.39
Riverine forest	9,463,849.50	5.26	4,046,700.00	3.13	51,605,307.00	10.41	22,066,200.00	7.52
Bushland	51,715,839.00	28.76	49,547,400.00	38.36	51,715,839.00	10.43	49,547,400.00	16.89
Total	179,823,937.00	100.00	129,151,700.00	100.00	495,958,594.00	100.00	293,341,400.00	100.00

Fig. 5.

The distribution of ESV of each LULC type based on the regional coefficient (1998, 2016).

In contrast, according to global estimates, grassland accounted for the highest ESV of approximately US$ 225.8 million (45.53%), followed by woodland and highland forest, which contributed approximately US$ 88.3 (17.82%) and US$ 78.4 million (15.81%), respectively, in 1998 (Fig. 4). In 1998, both bushland and riverine forests contributed comparably, with approximately US$ 51.7 (10.43%) and approximately US$ 51.6 million (10.41%), respectively (Table 6).

Fig. 4.

The distribution of ESV of each LULC type based on the global coefficient (1998, 2016).

The AWR experienced considerable ESV changes. Between 1998 and 2016, the regional ESV estimated an annual net loss of approximately US$ 50.6 million (28.18%). Over 18 years, the global ESV revealed an annual net decline of approximately US$ 202.6 million (40.85%). The ESV of grassland, riverine forest, and woodland has decreased greatly by 60.5%, 57.2%, and 37.7%, respectively, whereas bushland and highland forest have decreased by only 4.19% and 1.18%, respectively (Table 6).

3.3. Changes in individual ESVs

The ESVs provided by different ecosystem services are presented in Table 7. Like global and regional ESV, the individual ecosystem value assessment showed a dramatic decline in ESV. The total ESV has decreased from roughly US$ 374.5 to US$ 264.8 million. Provisioning and regulating services have declined from US$ 100 to US$ 70.6 million and US$ 242.4 to US$ 170 million, respectively. Further, supporting, and cultural services have declined from US$ 5.3 to US$ 3.2 million and from US$ 26.7 to US$ 20.8 million, respectively.

Table 7.

Ecosystem functions value coefficients of each LULC type in the AWR (The value is based on 2022 US$).

Ecosystem types	Ecosystem service values (US$ ha-1year-1 in each land use land cover
	Forest land		Grassland		Woodland		Bushland
Provisioning	1998	2016	1998	2016	1998	2016	1998	2016
Water supply	5,325,760.00	4,182,110.00	7,680,015.20	3,109,163.20	480,158.03	306,849.78	7,423,301.60	7,293,457.60
Food production	2,004,640.00	1,574,165.00	8,727,290.00	3,533,140.00	0.00	0.00	8,435,570.00	8,288,020.00
Flaw material	14,227,259.20	11,172,107.45	1,275,527.00	516,382.00	15,325,136.20	9,793,681.20	1,232,891.00	1,211,326.00
Biomass fuel	937,692.80	736,333.30	5,303,507.00	2,147,062.00	7,460,746.48	4,767,864.48	4,365,731.92	4,289,369.12
Genetic resources	9,804,185.60	7,698,841.60	671.33	271.78	0.00	0.00	648.89	637.54
Subtotal	32,299,537.60	25,363,557.35	22,987,010.53	9,306,018.98	23,266,040.71	14,868,395.46	21,458,143.41	21,082,810.26
Regulating
Water regulation	466,752.00	366,522.00	335,665.00	135,890.00	0.00	0.00	324,445.00	318,770.00
Water purification	8,233,984.00	6,465,824.00	6,693,831.43	2,709,918.38	28,117,521.96	17,968,782.96	6,470,082.19	6,356,911.34
Erosion control	10,771,200.00	8,458,200.00	3,752,063.37	1,518,978.42	2,967,443.16	1,896,374.16	0.00	0.00
Biological control	616,950.40	484,466.90	2,013,990.00	815,340.00	0.00	0.00	1,946,670.00	1,912,620.00
Biodiversity protection	7,868,960.00	6,179,185.00	3,536,566.44	1,431,737.04	1,268,593.04	810,707.04	3,418,352.52	3,358,560.72
Climate regulation	17,024,480.00	13,368,655.00	8,184,184.03	3,313,269.98	42,551,094.52	27,192,701.52	7,910,617.99	7,772,250.14
Gas control	661,830.40	519,709.40	161,790.53	65,498.98	0.00	0.00	156,382.49	153,647.14
Carbon sequestration	36,795,316.80	28,893,916.05	19,964,682.87	8,082,465.42	904,870.56	578,266.56	19,297,339.71	18,959,802.06
Subtotal	82,439,473.60	64,736,478.35	44,642,773.67	18,073,098.22	75,809,523.24	48,446,832.24	39,523,889.90	38,832,561.40
Supporting
Nutrient recycling	347,371.20	272,776.95	0.00	0.00	0.00	0.00	0.00	0.00
Pollination	1,444,836.80	1,134,573.55	2,148,256.00	869,696.00	0.00	0.00	0.00	0.00
Soil formation	97,539.20	76,593.70	469,931.00	190,246.00	0.00	0.00	0.00	0.00
Habitat/regulation	839,465.44	659,199.22	0.00	0.00	0.00	0.00	0.00	0.00
Subtotal	2,729,212.64	2,143,144.05	2,618,187.00	1,059,942.00	0.00	0.00	0.00	0.00
Cultural
Recreation	24,255,545.60	19,046,926.60	208,783.63	84,523.58	0.00	0.00	0.00	0.00
Tourism	2,206,300.80	1,732,521.30	0.00	0.00	0.00	0.00	0.00	0.00
Subtotal	26,461,846.40	20,779,447.90	208,783.63	84,523.58	0.00	0.00	0.00	0.00
Total	143,930,070.24	113,022,627.65	70,456,754.83	28,523,582.78	99,075,563.95	63,315,227.70	60,982,033.31	59,915,371.66

In the 1998 assessment, the value of regulating services accounted for about 64.7% of the total in the AWR, followed by provisioning services and cultural services, which accounted for approximately 26.7% and over 7.1%, respectively. Support services, on the other hand, have contributed to the lowest levels, accounting for about 1.4%. In the 2016 assessment, the regulating and supporting had slightly declined and the provisioning service remained unchanged, whereas the cultural service showed a slight increase.

Regulatory services, including carbon sequestration, accounted for approximately 20.55% in 1998 and 21.34% in 2016, respectively, followed by climate regulation and water purification services, which accounted for 20.21% and 19.51%; and 13.22% and 12.65% in 1998 and 2016, respectively. Between 1998 and 2016, the provision of flaw material services accounted for approximately 8.56% of all services.

3.4. Sensitivity analysis

The CS value exhibited a value less than one, indicating a relative decrease in the overall carbon sequestration potential. The woodland habitat demonstrated the capacity to enhance ecosystem services in 1998 and 2016, revealing values of 0.49 and 0.43 respectively. In 1998, the bushland habitat demonstrated the second-highest value of CS at 0.29, followed by the grassland habitat at 0.09. The highland and riverine forest habitats demonstrated the lowest CS values at 0.08 and 0.05, respectively. In 2016, the bushland and highland forest habitats showed CS values of 0.38 and 0.11, respectively, followed by the grassland and riverine forest habitats at 0.05 and 0.03, respectively, as presented in Table 8.

Table 8.

Change in estimated total ESV and coefficient of sensitivity (CS) after a 50% adjustment of ecosystem services valuation coefficient (VC) in the AWR.

Change of value coefficients	%	1998	%	2016
		CS		CS
Forest VC ± 50%	1.03	0.08	3.45	0.11
Woodland VC ± 50%	1.03	0.49	1.03	0.43
Grassland VC ± 50%	1.16	0.09	5.76	0.05
Riverine forest VC ± 50%	1.03	0.05	3.45	0.03
Bushland VC ± 50%	1.03	0.29	1.03	0.38

4. Discussions

4.1. LULC analysis

Human and livestock incursion is becoming a common phenomenon in most protected areas of East Africa [93]. Natural ecosystem replacement for other land uses such as farmlands, grazing lands, settlements, and urban areas is the prime factor of biodiversity loss, and land degradation [1]. Natural and socioeconomic factors, as well as spatiotemporal operations, influence LULC changes in the area. Accordingly, rapid LULC dynamics have been occurring in the area. According to Ref. [41], woodland and grassland have dramatically decreased, while bushland and cultivated land increased threefold and more than eightfold, respectively [41].

Likewise [42,46], discovered rapid habitat changes within the AWR and its surrounding grazing lands. This study found that important LULC changes occurred between 1998 and 2016. Bush encroachment, livestock incursion, and human interference have all potentially contributed to LULC in the area [42]. The AWR's grassland habitat is the most suitable for large mammals, but it is rapidly dwindling approximately 145 km² of grassland habitat has been lost.

In contrast, between 1998 and 2016, bushland habitat encroached on approximately 155 km² of the landscape; exotic plant species and the invasive Prosophis juliflora are thought to be the primary factors of bushland encroachment. The Prosopis was introduced in the late 1970s and early 1980s to aid in water and soil conservation in the Afar region [94]. Later, the species began invading croplands, grasslands, riverbanks, and roadsides; it has since been identified as the principal threat to biodiversity and livestock production [95].

4.2. ESVs assessments

The most important factor contributing to the differences in the findings was the use of different ESVs [68]. Due to important differences in ESV US$ ha⁻¹ yr⁻¹ for various land cover types, the estimated ESV also showed important differences in the AWR during the study periods. The rapid encroachment of bushland on grassland habitat is the major contributor to total ESV change. The boundary extension towards Mount Asebot in 2014, on the other hand, has contributed to the legal protection of the highland forest under AWR management and the nearby monastery. The ESV of grassland, riverine forests, and woodland has drastically decreased, whereas bushland and highland forests have increased ESV. According to Ref. [54], the changes in ESVs initiated by LULC dynamics have resulted in a loss of approximately US$ 112 to 1091 million in the Afar region, which contains an important portion of the AWR.

Furthermore [68], reported a slight increase in total ESVs in the Abaya-Chamo basin, but a decrease in total ESV from natural ecosystems (for example, Nech Sar, which is part of the basin). Similar results were reported in previous works by Ref. [56] in the Shashemene Munisa forest [65]in the Andessa watershed of Blue Nile; [96] in the central highlands of Ethiopia; [97] in the Winkie watershed of Omo Gibe basin; and [54] in the Afar region who reported about the declined ESVs for the grazing land, forestland, shrubland, and water body. The annual ESVs of the AWR were estimated to be US$ 84.48 million [60]. The 2015 estimate was lower than the global, regional, and distinct 19 ecosystem services estimate because the 2015 study only evaluated nine ecosystem services, resulting in lower AWR ESVs.

4.3. Limitations of the study

The evaluation of LULC alterations could be restricted by the caliber and determination of the data employed [68]. The data utilized in this investigation are derived from assorted origins, which might be the primary constraint of the research, as it is almost impossible to avoid uncertainties in LULC transformations simulated due to several factors [68]. As a result of the restricted data sources, it was not feasible to measure the effect of forthcoming land-use shifts as well as other significant obstacles, including climate change, on the landscape ecosystem.

When employing this research in policymaking and decision-making methods, it is critical to understand the origins of uncertainty in LULC projections. Moreover, the complexity of assessing ecological changes using land cover types and the corresponding "biome category," to which the assigned ESVs belong in a more comprehensive context, as well as the significant variation that can occur when applied in the local context, is another limitation of this study.

5. Conclusion

This study investigates the alterations in ESVs as a response to LULC dynamics within Ethiopia's protected area system, with a particular emphasis on AWR. The ESVs in AWR were significantly impacted by LULC changes between the years 1998 and 2016. The reduction of grassland and bushland encroachment played a prominent role in ESV changes. Despite the decision-makers acknowledgment of the value of AWR in the conservation of endangered wildlife species, a substantial anthropogenic obstruction to improving the reserve's management persists. Therefore, the results of this study could be utilized to augment decision-maker's comprehension of protected areas in general and more specifically of the AWR, as well as to inform future planning and management initiatives aimed at the conservation of AWR's pristine rangeland. These interventions would, in turn, enhance ecosystem services and improve the livelihoods of surrounding communities.

Availability of data and material/data availability

The author wants to declare that data can be submitted at whatever time based on request. The data used for the current study are available from the corresponding author upon reasonable request.

Code availability (software application or custom code)

Not applicable.

CRediT authorship contribution statement

Simeneh Admasu: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Resources, Data curation, Conceptualization, Formal analysis, Investigation, Methodology, Project administration.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.