Watershed management intervention on land use land cover change and food security improvement among smallholder farmers in Qarsa Woreda, East Hararge zone, Ethiopia

ABSTRACT

Natural resource bases are the basic foundations of food security. However, they are facing problems both in quantity and quality, leading to decreased land productivity and societal issues. Watershed management is the most effective technique for managing and utilising these resources while improving food security. The study evaluated the effectiveness of the watershed management approach using Arc GIS 10.5 to track changes in land use cover for the years 2009, 2014, and 2020 and surveyed 337 households in three different Programmes. The findings showed six land-use land cover areas, where the proportion of farmland and bare land decreased while the proportion of settlements and shrub land increased in the three micro-watersheds. Population pressure and exotic species' invasion were significant factors contributing to the decline in farmland, while a decrease in bare land and an increase in shrub land were the results of successful watershed management interventions. Survey results showed that approximately 93.4% and 84.9% of respondents reported reduced soil erosion and deforestation, respectively. Further, the integrated watershed approach implemented in the area improved water availability, reduced dependency on single crops, and increased off-farm activity. Thus, scaling up an integrated watershed approach can enhance food security and protect natural resources.

KEYWORDS:

• Natural resources
• production, and productivity
• watershed approach

Introduction

There is growing consensus in Ethiopia that improving the country's natural resource base is crucial for ensuring food security and self-sufficiency (FAO 2017; Bakala and Asfaw 2020). They serve as the foundation of food security, either directly by generating the necessary food or indirectly by producing goods and services that may be exchanged for food (Wani et al. 2010). As described by Gashu and Muchie (2018) natural resources are the supreme gifts without which life in any form cannot exist on planet Earth. As several research studies have also shown, managing the natural resource base is essential for improving production and enhancing human well-being both now and in the future (Gashu and Muchie 2018; Ewunetu et al. 2021). However, these resource bases are facing serious challenges both in terms of quantity (per capita) and quality, resulting in decreased land productivity and increased social problems (Bazezew and Worku 2018; Meyfroidt et al. 2019). Feeding a growing population will become increasingly difficult under the current management practices, needing prompt remedial action (FAO 2017).

In this respect, it is found that watershed management is among the best tools to ensure sustainable utilisation and management of natural resources to obtain the desired goods and services without adversely affecting environmental resources (Worku and Sangharsh 2015; Reddy et al. 2017). Furthermore, studies conducted over the past 50 years in India and China on watershed management showed remarkable successes in improving the basic conditions of agricultural production and promoting the rural economy, especially in rain-fed and drought-prone regions (Wani et al. 2010). In Ethiopia, a study conducted by Gebregziabher et al. (2016) in three regions (Oromia, Amhara, and Tigray) showed that watershed management has increased farm incomes and food security by an average of 50 and 56 percent, respectively. Additionally, Ethiopia's Climate Resilient Green Economy Policy and Strategy issued in 2011 reiterates the significance of watershed management as a part of the nation's economic development to enhance the condition of the natural resources and meet the demands of the population for sustainable livelihoods.

Despite these facts, the effectiveness of an integrated watershed management approach as a method of reducing poverty and improving food security was not widely recognised among the various actors (Worku and Sangharsh 2015; Mengistu and Assefa 2019). As a result, attention given to the integrated assessment of the biophysical and socioeconomic components of watershed management for improving food security was minimal. However, the management of watersheds and food security are closely related. For instance, several issues that cause food insecurity are linked to the degradation of the natural resource bases, including land degradation, drought, floods, poverty, conflict, poor health, and ineffective governance (Adimassu et al. 2017; Mekuriaw 2017; Mena 2018). Numerous research studies have indicated that managing the natural resource base will increase the sustainability of water availability, production from agriculture, income, and environmental, and ecological status and also contribute to climate change mitigation (Legesse et al. 2018; Gopa et al. 2021). However, only a few studies have examined how comprehensive watershed management practices contribute to ensuring food security.

Furthermore, given the long history of watershed management in the country in general and the study area in particular, there have only been a few attempts to evaluate the biophysical, socioeconomic, and environmental components of watershed management practices to ensure food security. Many past studies have attempted to primarily focus on specific criteria, ignoring the other components of the watershed development (Desta et al. 2005; Gebregziabher et al. 2016). In addition, Bekele et al. (2018) and Worku and Sangharsh (2015) reported that there was a lack of clear understanding to guide the integration of multiple actors in watershed management to alleviate land degradation while improving food security. A comprehensive study on the approach's effectiveness in improving community food security while improving natural resources is so important. Thus, the objective of this research was to assess the contribution of biophysical, and environmental components of watershed management practices to assuring food security under different programme approaches including the Sustainable Land Management Project (SLMPII), the Productive Safety Net Program (PSNP PW), and the Free Community Mass Mobilization programme in Qarsa Woreda of the East Hararghe zone of Oromia region.

Materials and methods

Conceptual framework of the study

The multiple links between watershed management and food security are expressed in various scientific literature (Wang et al. 2016; Bakala and Asfaw 2020). These research works vividly show that the concepts of food security and sustainable watershed management are closely interrelated. Watershed management is the most effective method for managing natural resources, and it has either directly or indirectly enabled the four pillars of food security – accessibility, sustainability, and utilisation – of food. In addition, Guiné et al. (2021) and Moroda et al. (2018) confirmed this idea by explaining that food security in Ethiopia is affected by a variety of factors including the environment, availability and quality of natural resources, including water resources as well as economic factors which among other things include, availability of oxen, land size, social and political/policy frameworks including the scope of the level of participation of stakeholders in decision-making processes.

According to FAO (2017), watershed management interventions also address the basic livelihood needs of poor people- such as drinking water supply, savings and credit, transportation, communications, off-farm income generation, and access to health and education services. As a result, using a watershed management approach is one of the best strategies for organising land use and other resource holistically (Anantha and Wani 2016; Gopa et al. 2021). Watershed management has a positive impact on a variety of biophysical, environmental, and socioeconomic aspects, including income, employment, asset creation, health, education, and the mode of use of energy resources all of which contribute directly or indirectly to the four dimensions of food security (FAO 2017; Teressa 2020). The study used biophysical, social, and economic sustainability outcomes such as reduced runoff, reduced downward watershed flood, enhanced soil fertility, enhanced vegetation, improved crop yields and farm income, and enhanced food supplements for livestock as tools for investigating the effectiveness of watershed management as a sustainable approach for the promotion of food security, as shown in Figure 1.

Figure 1. Conceptual framework of the study modified from Vishnudas, (2006) and Mawhorter, (2010) and from experience.

Description of the study area

The study was conducted in Qarsa Woreda of East Hararghe zone, Oromia region of Ethiopia. Geographically, the district lies between 90 17’ and 90 29’N latitudes and 410 12’ and 410 56’E longitudes to the west of the zonal capital, Harar town. The woreda is one of the food insecure in the East Hararghe zone. The livelihood of the rural inhabitants is closely linked to the use of land resources for food production, energy sources, shelter constriction, and so on. Watershed management strategies have been applied in the area for a very long period through different initiatives and programmes. However, the woreda continued to be the most affected by difficulties with chronic food and nutrition security, soil erosion, plant loss, declining soil fertility, and water stress. This study focused on three different programme approaches: Free Mass Mobilization, Sustainable Land Management Program II, and the Productive Safety Net Public Work which are adjacent to each other (Figure 2). All the programmes are using watershed management as one of the main tools for improving food security and reducing poverty, but the experiences and programme approaches have varied. Free Mass Mobilization is the government's regular programme and has no external funding support. Sustainable Land Management Program II and Productive Safety Net Program are funded by the World Bank.

Figure 2. Map of the Study Area.

The description and characteristics of each selected watershed are presented in (Table 1). Agriculture is the main economic activity of the rural residents with mixed crop-livestock production systems on a subsistence level. The rain-fed production system is the most dominant and practiced by the majority of the farmers. Maize (Zea mays L.) and Sorghum (Sorghum bicolor) are produced as staple crops while Khat (Catha edulis) and coffee (Coffea arabica) as important cash crops. Livestock is also an important source of food and household income.

Table 1. List of micro watershed management systems covered by the study.

No	Name of micro watershed	Implemented by	Agro-Ecology	Altitude	Soil type	Area in Ha	Population
							HH	Sample size
1	Andhura Kosum	Free Mass Mobilization	H/L	2000–2700	Luvisols, Lithosols, Cambisols	759.5	734	112
2	Baraka	PNSP	H/L	1900–2700	Luvisols, Lithosols, Cambisols	565.7	639	107
3	Burqa Water	SLMP	H/L	2100–2700	Luvisols, Lithosols, Cambisols	853.1	765	118
						2178.3	2138	337

PNSP = (Productive Safety Net Program), SLMP = (Sustainable Land Management Program).

Based on 20 years (2001–2020) of data obtained from Ethiopian National Meteorology Agency Qarsa Woreda receives an average annual rainfall of 1225 mm/yr. The annual mean minimum and mean maximum temperatures were 12.5 and 26.6 ⁰C respectively. The woreda is characterised by a bimodal rainfall pattern with an erratic distribution. The small rainy season Belg(Arfasa) occurs from March to May, while the main rainy Kiremt (Gana) season occurs from June to September with a dry season from October to February (Figure 3).

Figure 3. Mean annual temperature and rainfall of Qarsa woreda (2000–2020).

Data requirements, sources, and methods of collection

Both primary and secondary data were employed to meet the study's objectives. The primary data was gathered through focus group discussions, key informant interviews, household surveys, and field observation. The secondary data, such as satellite imagery and information on previous conservation initiatives, was acquired from relevant websites and reports. The primary data from the field survey were supplemented with secondary data to substantiate the study.

Data sampling technique

This study used a three-stage random sampling technique to select the micro watershed and household heads. The woreda was purposively selected due to the presence of Free Mass Mobilization, Sustainable Land Management Program II, and Productive Safety Net Program in the woreda. The selection of the micro watershed was based specific set of criteria such as the historical similarity before and after watershed management interventions, similar land use systems and soil and water conservation practices used, and sites within adjusting to each other. Accordingly, one micro watershed from each programme approach was selected. The households of the micro watershed were considered for the population of the study. The description and characteristics of each selected watershed are presented in Table 1. Following the identification of the total households in the micro watershed, we determined the sample size by using the sample size determination equation that takes into account the desired confidence level (95 percent), error margin (5 percent), and non-response rate (10 percent) where the exact number of households in each micro watershed is known by using (Kothari 2004) sample size correction calculator as indicated hereunder. (1) $\mathrm{S}={\mathrm{Z}}^2{}^{\ast }{P}^{\ast }(1\text{-}\mathrm{P})$(1) C2

Where: Z = Z-value (1.96) for 95 confidence level.

  P = is the percentage picking a choice, expressed as a decimal (0.5)

  C = is the confidence interval expressed as a decimal (0.05 = ±0.05)

Subsequently, the actual sample size for the study area was determined as: (2) $S{S}_{kp}={\displaystyle \frac{S}{1+{\displaystyle \frac{S-1}{P_2}}}}$(2) Where: SSkp is the sample size for the known population size

   S is the sample size for the unknown population calculated using Equation (1).

   P_k is the known population size from which the sample size is calculated.

In addition to data collected from household surveys, focus group discussions (each comprising 7–12 participants) were selected purposively as they were expected to be experienced and knowledgeable about the issues under investigation. Accordingly, 5 FGDs were conducted one at the watershed level (kebele level) 1 at the district, and 1 at the zone level based on the prepared checklists.

Satellite data source and processing

To analyze the land use and land cover of the study watershed, multi-spectral, multi-temporal LANDSAT satellite data of East Hararghe of Qarsa Woreda were acquired for three years namely, 2009, 2014, and 2020. All LANDSAT images have been taken from USGS which were from January to ensure uniformity and low cloud cover images. The satellite images were brought to the Universal Transverse Mercator (UTM) projection in zone 37 N and were used to investigate land cover changes in the study area. In addition senile image was also used for its high spatial resolution for identification different land use types like bodies detection and settlements. The selection of each satellite image was done with various objectives. The year 2009 was selected to trace back the status of the watershed before the implementation of the watershed management projects. The year 2014 was selected because it is the time when the watershed management intervention started. The year 2020 was selected to see the impact of the watershed management strategies on the biophysical change after the implementation of the watersheds. Each satellite image is presented in detail in Table 2. Global positioning system (GPS) points collected during field observation were used to collect GCP (ground control point) to successfully undertake the image classification.

Table 2. Description of Landsat image.

No.	Date production	Pass/Row	Scale	Source
1	January, 2009	166/54	30 m	Earth Explorer USGS
2	January, 2014	166/54	30 m	Earth Explorer USGS
3	January 2020	166/54	10 m	Earth Explorer USGS

Source:(Hassan et al. 2016;Mzava et al. 2019).

Determination of land use classification system

The land use land classes of interest were carefully selected and established using information from key informants, field observation, expert opinions, a reconnaissance survey of the existing LULC distribution in the study, and the use of related previous studies. Accordingly, six main land-use and land-cover classifications have been identified for the analysis of changes in land use and land cover (Table 3).

Table 3. Description of land use and land cover categories considered in image classification.

LULC	Description
Agricultural land	Land areas mainly under rain-fed farming of annual crops growing and perennial crops. Major, crops grown include cereals (maize, wheat, barley, etc.), spices, and cash crops (coffee/chat).
Forest land	Areas covered by large trees with open canopies including natural forests and plantations. covered by trees with a canopy cover of 40–65 percent.
Shrub/bushland	Covered by multi-stemmed scrubs, sparse small trees, and shrubby undergrowth. Found in the escarpments, on steep slopes (either at early years of enclosure or area closure)
Bare land	Areas with no or very little vegetation cover and are characterised by the shallow and rocky surface along the flooding area of the local stream valleys, over gentle and steep mountain slopes
Rural settlements	Areas occupied by residential houses of rural communities, buildings, and small towns
Waterbody	Areas covered by open waters (mainly lakes, reservoirs, streams, canals, and others)

Source: and Authors field survey and observation.

Accuracy assessment

Accuracy assessments were performed for classified images from 2009, 2014, and 2020. Parallel to the remote sensing work, other data sources, such as GPS data from field visits, topographic maps, and raw images, were collected (Congalton & Green. 2009). During the field visit, 36 GCPs were collected using GPS and Google Earth, and classification accuracy was assessed using error matrix analysis, overall accuracy, and Kappa coefficient analysis. The error matrix expresses the accuracy of the user and producer, whereas the Kappa coefficient agrees with the interpreters (Foody 2022). The accuracy of the producer and user was calculated using a confused matrix that included the overall accuracy and Kappa coefficient (Congalton 2001). All output maps had to be at least 85% accurate. To detect changes in land use and land cover, a visual comparison of features and a confusion matrix analysis were performed. The areas that were converted from one class to any of the other classes were computed, and the directions of the changes were determined.

Overall accuracy:

According to Bhatta (2008), an error matrix is ⁣⁣one way to express the accuracy of classifications (confusion matrix or contingency table). To use the error matrix the ground control point must collect as a viable sample based on which an error matrix is formed. Generally, 6 samples for control points as data sampling were collected, and the relationship between known reference data (ground data) was calculated as 83.3%.

Kappa coefficient:

In the classification process, where pixels are randomly assigned to classes will produce a percentage correct value. The resulting Kappa measure compensates for chance agreement in the classification and provides a measure of how much better the classification performed in comparison to the probability of randomly assigning pixels to their correct categories.

Therefore, it is calculated as follows: $k={\displaystyle \frac{N{\sum }_{i=1}^{n}mi,i-{\sum }_{i=1}^n(G_i{C}_i)}{N^2-{\sum }_{i=1}^n(G_i{C}_i)}}$When:-

i the class number

N is the total number of classified values compared to truth values

m_{i, I} the number of values belonging to the truth class

i that have also been classified as class I (i.e. values found along the diagonal of the confusion matrix)

C_i is the total number of predicted values belonging to class i

G_i is the total number of truth values belonging to class i

Accuracy assessment in this study revealed a kappa coefficient of 76.66%, 76.66%, and 80.00% for 2009, 2014, and 2020, respectively. The overall accuracies of 80.55%, 80.55 and 83.33% were calculated for 2009, 2014, and 2020, respectively

Image classification and detection of changes

Arc Map 10.7 was used to create land use maps for 2009, 2014, and 2020, which were then analyzed as follows: (a) supervised classification allegorising tool; (b) displaying all the different classes in the same layer; and (c) calculating the area of each class. Then, changes in land use patterns were detected using a land use table that included areas of various classes. While the three components have different impacts, the reasons are unclear: why have higher levels of social and environmental impacts failed to translate into economic impacts?

Data analysis methods

ArcGIS 10.5 and ERDAS IMAGINE software were used for image classification, land use, land cover change detection, and mapping. Data from interviews, focus groups, field observations, and secondary sources were edited, coded, and entered into SPSS 26 and Stata. then, the biophysical characteristics of the watershed and information from household surveys be used for interstation.

Results and discussion

Land use and land cover change detection over the period 2009, 2014, and 2020

Figures 4–6 show land use and land cover maps generated from ground truth and classified Landsat images for three time periods (2009, 2014, and 2020), demonstrating that land use changed from one land use type to another in an increasing or decreasing manner. Better land use pattern is one of the important objectives of watershed management.

Figure 4. Land use the land cover of Burqa Water (2009–2020).

Figure 5. Land use land cover of Barka (2009–2020).

Figure 6. Land use land cover of Adhura Kusum (2009–2020).

During the study period, six different categories of land use land cover were identified: settlements, water bodies, bare land, agricultural/farmland, shrub and brush land, and forest land. Farmland was the most dominant land use type in all three micro watersheds, accounting for around 80.24 percent of Burka water, 77.75 percent of Andhra Kosum, and 74.59 percent of Barka micro watersheds. Interestingly, the proportion of farmland and bare land decreased in all three intervention areas by 0.48 and 0.4 percent in Burka Water, 0.7 and 0.55 percent in AduraKosum, and 0.76 and 0.57 percent in Baraka micro watersheds. Settlement area and shrub land, on the other hand, increased in all three micro watersheds by 0.41 and 0.49 percent, 0.63 and 0.62 percent, and 0.76 and 0.68 percent in Burka water, Andura Kosum, and Barka micro watersheds, respectively. The decrease in bare land in both micro watersheds indicates the effectiveness of watershed management interventions in restoring degraded land. This result aligns with other studies by Assefa and Singh (2018) and Tesfay et al. (2023). However, the decrease in farm / agricultural land is contrary to the general trend of previous land use and land cover change studies in Ethiopia and Assefa and Singh (2018) but consistent with Tesfay et al. (2023). Key informants interviewed also confirmed that population pressure is the main factor contributing to the reduction of farmland, and it may remain a problem even in the future despite the positive impact of the watershed intervention on reducing bare land. In addition, during the focus group discussion, they explained that the spread of invasive alien plant species such as Striga and Parthenium hysterophorus on agricultural land was another reason for the decrease in farmland. A study on the impact of an invasive alien weed species in Eastern Hararghe also confirmed that the invasion of farmland by invasive alien plant species reduced farmland for crop production. A 90-year-old farmer from Adhura Kosum shared a story that illustrates the impact of population pressure on land ownership.

He explained that he shared his 3-hectare land with his four male children, who then divided their share with their grandchildren. The children of his grandchildren have no more land to share with the next generation because they only receive a piece of land on which they can construct their homes. The opportunity to inherit land in his family is almost impossible.

This is in line with the general trend of central statistical agency population projections which indicated the average land holding size per holder has in general been decreasing due to population growth and the degradation of agricultural land. The findings also highlighted that sustaining food security in the area through agricultural land is the most pressing concern now and in the future and that establishing an integrated watershed Management approach is an essential strategy in the area.

Land use land cover inter-class conversion and change path

The conversion matrix displaying the LU LC change gained, lost, substituted, and net persisted values for the years (2009–2014) and (2015–2020) is shown in Tables 4 and 5. From(2009 to 2014) LU LC change conversion, the highest decrease occurred on bare land, followed by cultivated land, while settlement areas, bushland, and water bodies showed an increasing trend. According to LU LC conversion, between (2014–2020) the highest decrease occurred in barren lands followed by cultivated land, whereas settlement areas, bushland, water bodies, and forest land showed an increasing trend. The key informant interviews, focus group discussions, and image interpretation confirmed that bare land is decreasing and shrub land and waterbody are increasing. Further, practical field observations also revealed that the watershed management practice resulted in a biophysical change in all of the study micro watersheds. The land use and land cover change study revealed the success of integrated watershed management practices in restoring degraded land, which contributed to increased water availability (Table 6). A similar study by Tsegaye et al. (2014) also reported the impact of watershed management on changes in land use and land cover and increased production.

Table 4. Matrix for land use and land cover changes for 2009–2014 in Hectares (ha).

LULC in 2009(ha)		LULC in 2014(ha)
		Bare Land	Settlement	Shrub Land	Farm Land	Forest Land	Water Body	Grand Total
LULC in 2009(ha)	Bare Land	105.86		34.22		1.30		141.38
	Settlement		95.81					95.81
	Shrub Land			28.32	91.75	0.86		120.93
	Farm Land		14.31	72.90	1600.34			1687.56
	Forest Land			0.95	13.43	64.26		78.64
	Water Body					2.56	51.50	54.06
	Grand Total	105.86	110.12	136.39	1705.53	68.98	51.50	2178.38

Table 5. Matrix for land use and land cover changes for 2015–2020 in Hectares (ha).

LULC in 2014(ha)		LULC in 2020(ha)
		Bare Land	Settlement	Shrub Land	Farm Land	Forest Land	Water Body	Grand Total
LULC in 2014(ha)	Bare Land	91.15	3.95	7.51	0.88		0.03	103.51
	Settlement		106.97					106.97
	Shrub Land	0.05	0.07	130.15	4.85	0.10	0.00	135.21
	Farm Land	0.70	11.70	8.23	1696.74	0.94	0.48	1718.79
	Forest Land			0.05	1.03278_0.055	64.97		65.02
	Water Body	0.00	0.03	0.08	0.16		48.62	48.88
	Grand Total	91.98	121.69	146.09	1701.02	66.02	49.13	2178.38

LULC = land use land cover change.

Table 6. Household response on effectiveness of watershed implementation on Biophysical land rehabilitation

No	Status after Watershed Management Interventions	Increased		Decreased		No change		Difficult to tell		Worsen
		Freq.	%	Freq.	%	Freq.	%	Freq.	%	Freq.	%
1	Reduced soil erosion	5	1.5	313	93.4	12	3.6	4	1.2	1	0.3
2	Reduced Deforestation	25	8.6	248	84.9	16	5.5	1	0.3	2	0.7
3	Overgrazing	12	3.7	283	86.5	27	8.3	4	1.2	1	0.3
4	Gully formation	23	7.1	256	79.3	40	12.4	4	1.2	0	0.0
5	Plough steep slopes	23	7.1	237	73.4	52	16.1	8	2.5	3	0.9
6	Cutting trees for fuel wood	44	13.6	224	69.1	44	13.6	11	3.4	1	0.3

Observed changes in ecological status as perceived by farmers after the watershed management intervention

According to Key Informant interviewers, before the watershed management intervention, there was a severe problem of soil erosion and floods flowing from the upper slopes, destroying many individuals’ farmland. The massive removal of soil from farm and pasture areas has resulted in crop production problems and livestock feed shortages. A study conducted by Tsegaye et al. (2014) in the surrounding area also confirmed that the problem of flooding from the sloppy area has destroyed many farmlands and resulted in the disappearance of various water bodies like Lake Alamaya. in the area. However, as stated by key informant interviewers following the watershed management intervention, various physical conservation structures have been constructed in the area, resulting in several noticeable improvements such as the restoration of degraded areas, reduced soil erosion, and the reemergence of previously disappeared water bodies.

Similarly, during the Focus Group Discussion, participants from the three micro watersheds confirmed that the problem of soil erosion, sedimentation, and flooding had been reduced, and as a result, crop and pasture productivity had improved. The practical field visit also confirmed the visible changes as a result of the watershed management intervention in the restoration of degraded land and the reemergence of previously disappeared water bodies. This is congruent with studies by Legesse et al.(2018), and Gebregziabher et al. (2016) which reported that watershed management practices reduced the rate of soil erosion, and sedimentation in the downstream, and improved soil moisture while increasing productivity.

Similarly, the household survey confirmed both the idea raised by the key informant and the focus group discussion as indicated in the table (Table 6). Almost all of the households interviewed in the three micro watersheds stated that the soil and water conservation practices, as well as the enclosed area, enhanced their livestock feed. About 93.4 percent of surveyed households said that soil erosion had decreased after the watershed intervention, 3.6 percent reported no change, and 1.5 percent reported that soil erosion was increasing. Furthermore, 84.9 percent of the respondents said that deforestation has been reduced. 5.5 percent answered there had been no change, and 8.6 percent stated that deforestation was increasing. Similar studies by Tadele and Dananto (2018); Adimassu and Langan (2016); and Haregeweyn et al. (2012) also found that watershed management has been proven to preserve sustainable biological conditions by restricting surface runoff generation, which results in an incremental rise in groundwater that is similar to the finding in Table 6. A study conducted by Munro et al. (2008) in Tigrayi's central plateau also found watersheds effective in controlling soil erosion and increasing land productivity. However, the key informant stated that poor policy implementation and lack of clear roles and responsibilities among different stakeholders are still major challenges for sustainable watershed management in the area.

Ecological and economic benefits that contribute to increased food security

Attaining food security through its four pillars is heavily dependent on the effectiveness of land and water management practices as well as the success of other sectors (Diriba 2020). In line with this, the results of the study indicate that watershed management practice contributes to the four dimensions of food security: availability, access, utilisation, and stability. About 89.5% of the households reported that after the watershed management intervention, their dependency on a single crop was reduced as a result of diverse possibilities. This is in line with the findings of the study conducted in Tigray by Adimassu et al. (2017) which examined how farmers were able to grow a variety of fruits, vegetables, trees, and forages after rehabilitating degraded land. The interviewed households have also confirmed the effectiveness of watershed management on their socioeconomic improvement in terms of enhanced agricultural input, credit services, increases in off-farm income, reduction in post-harvest loss, irrigation development, health, and veterinary service. Table 7 also demonstrates the environmental and economic improvements that the respondent households witnessed after the watershed management intervention in the three micro watersheds.

Table 7. Household responses to the effectiveness of Watershed Management Practices for socioeconomic improvement.

No	Improvement after watershed management	Increase		Decrease		No change
		Freq.	%	Freq.	%	Freq.	%
1	Improvement of agricultural and rural credit after IWSM implementation	220	65.3	20	5.9	97	28.8
2	Irrigation practices after IWSM implementation	260	77.2	7	2.1	67	19.9
3	Pest infestation after IWSM implementation	238	70.6	10	3.0	89	26.4
4	Dependence on single harvest after IWSM implementation	21	6.2	266	78.9	50	14.8
5	Off-farm income improvement after IWSM implementation	294	87.2	6	1.8	37	11.0
6	Post-harvest losses after IWSM implementation	7	2.1	268	79.5	62	18.4
7	Health /malaria problem after IWSM implementation	11	3.3	295	87.5	31	9.2

IWSM = (Integrated Watershed Management).

Practical field observations also revealed the potential of watershed management, which involves combining various activities on a single piece of land to contribute to food security. The majority of the households interviewed in the three micro watersheds also reported that watershed management interventions had enabled the restoration of degraded which resulted in a significant improvement in land productivity and thereby ensured their food security. A similar study on watershed management by Mekuriaw (2017) and Gashaw (2015) also confirmed that physical soil and water conservation with agronomic practices are essential for improving both the provisioning and management of ecosystem services, as well as food security. Studies by Mulugeta et al. (2018), Assefa and Singh (2018), and Mohammed (2015) also confirmed that watershed management activities help to increase household income, and access to water for irrigation, human, and livestock.

The focus group and key informants also revealed that before the implementation of the watershed intervention in the areas, women and children had to travel long distances to fetch water. However, following the development of the watershed, small springs emerged, as a result of which the local people began to fetch water from a short distance close to their residences. They also confirmed that soil conservation and water harvesting structures in the watershed helped increase water availability for cattle and other domestic uses, consequently improving the perennial supply of water from the streams. The majority of the interviewed households also revealed that after the watershed management intervention, they had noticed changes in the socioeconomic and environmental conditions. A key informant interview and focus group discussion further demonstrated that the physical and biological conservation measures implemented had resulted in the rehabilitation and reclamation of severely degraded areas, which resulted in notable improvements in land productivity. Table 8 also confirms this.

Table 8. Household Repose on effectiveness of Watershed Management on access to water and livestock feed.

Status of Watershed Management after Interventions	Increased		Decreased		No change		Difficult to tell		Worsen
	Freq.	%	Freq.	%	Freq.	%	Freq.	%	Freq.	%
Access to livestock feed	197	60.2	52	15.9	71	21.7	4	1.2	3	0.9
Access to potable water	207	63.9	35	10.8	74	22.8	7	2.2	1	0.3
Access to irrigation water	235	72.3	29	8.9	56	17.2	4	1.2	1	0.3
Crop productivity	236	72.6	26	8	53	16.3	10	3.1	0	0
Livestock productivity	188	57.5	30	9.2	86	26.3	23	7	0	0

Secondary data from the three micro-watersheds also confirmed that the irrigated area has increased by 0.72 percent, in the Burka Water 0.69 percent in Adhura Kosum, and 1.5 percent in Barak after the watershed intervention. Interviewed agricultural experts also confirmed that watershed management practices have assisted crop and livestock production and productivity. This will in turn improve household food security. The findings are consistent with those of Mekuriaw (2017) and Gashaw (2015), who also reported that watershed management programmes increased crop yields, reduced run-off and soil loss, increased groundwater, and reduced poverty. Yet, the main challenge is that neither government officials nor practitioners seem to understand the connections between food security and integrated watershed management practices. Therefore, the local community, extension agents, experts, and policymakers must give attention to the watershed management approach to make the watershed a viable solution to the problem of food insecurity.

Household response on food security status from own farm after and before watershed management intervention

In the context of food security, smallholder farm productivity is crucial for both food production and household consumption levels. In this section, we compared the number of months people experienced food shortages before and after the intervention to learn more about the effects of the biophysical and socioeconomic changes brought about by watershed management on rural household food security in the study area. The PSNP's experience, which involved transfers for six months at times of shortages of food, was used as a benchmark.

Before the implementation of watershed management practices, only, around 16% of PNSP PW, 5% of SLMPII 23% of mass mobilisation reported experiencing food shortages for one to three months. However, following the implementation of watershed management, the proportion of respondents reporting food shortages for one to three months changed to 38.1%, 49.5%, and 52.7%, respectively, for the PNSP, SLMPII, and mass mobilisation interventions. Additionally, following watershed management practices, the percentage of PNSP PW who responded to food shortages lasting longer than 10 months decreased from 16.9% to 2.56%. Similar results were also observed in SLMP II and Free Mass Mobilization, where 35% and 33% of respondents, respectively, showed the complete transformation from experiencing 7–9 month food shortages to 3–6 months after watershed management intervention. Statistical analysis of the chi-square test also showed a significant difference at P-Value  = 000 in both the pre-and-pre-and-pre- and post-watershed management intervention. However, productive Safety Net and SLMP programme participant farmers showed a significant improvement in their food security status as compared with free mass mobilisation programme participants.

Our findings suggest that watershed management improves the food shortage of the respondent households by 62% as compared to before the watershed management (Table 9). Similarly study by Berhane et al. (2014) in PNSP public reported that watershed management has improved food security by 1.29 months. Studies conducted in different parts of the country also confirmed the effectiveness of watershed management interventions in enhancing food availability, water access, and reliable income (Gashaw 2015; Mekuriaw 2017; Worku et al. 2018). However, challenges still exist about the status of food security as both programme approaches are focused on physical conservation rather than comprehensive watershed management. Therefore, if food security is to be sustained in the long run, equal understanding and knowledge of watershed management needs to be implemented in all three programme approaches.

Table 9. The number of months of food shortage before and after implementing integrated watershed management.

No of Month	Baraka WS_PNSP N = (118)		BurkaWater_SLMP N = (107)		Adhura kosum-Regular N = (112)		Total N = (337)
	Before %	After %	Before %	After %	Before %	After %	Before %	After %
1–3 month	(19) 16.1	(45) 38.1	(6) 5.6	(53) 49.5	(26) 23.2	(59) 52.7	(51) 15.1	(157) 46.6
3–6 month	(52) 44.1	(39) 33.1	(63) 58.9	(54) 50.5	(48) 42.9	(53) 47.3	(163) 48.4	(146) 43.2
7–9 month	(27) 22.9	(34) 26.3	(38) 35.5		(37) 33.0		(102) 30.3	(34) 10.1
Above 10 month	(20) 16.9	(3) 2.5			(1) 0.9		(21) 6.2%	(3) 0.89

Chi-Square = 51.846 P-Value = 0.00 before 56.616 P-value = 0.00 after.

Conclusions

The findings of the study in both three micro watersheds on land-use land cover change and Household survey have indicated that watershed management interventions have yielded tangible results. The land use dynamics between 2014 and 2020 showed a decrease in bare land area and an increase in shrub lands, forest land, and water bodies. The household survey also confirmed that watershed management intervention has played an important role in reducing soil erosion, improving moisture retention and soil productivity, and allowing farmers to diversify their production alternatives and incomes. This would indicate how much a watershed management approach improves degraded agricultural land, thereby attributing to increased production and productivity, water availability, and fodder security for the community while mitigating the effects of drought. However, the watershed approach is still more focused on physical conservation rather than comprehensive watershed management. Therefore, if food security is to be sustained in the long run, it is crucial to develop the linkages between the environmental, social, and economic issues to manage watersheds effectively.

Acknowledgments

First, let us give thanks to God for this day and time. Second, we thank all the respondents’ farmers and experts for their cooperation in providing the data for this study

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data will be made available on request.

Additional information

Funding

This research did not receive any specific grant from funding agencies.

Notes on contributors

Tena Gobena

Tena Gobena is PhD student at the Center for Food Security Studies, College of Development Studies, Addis Ababa University. He obtained his first degrees from Haromaya University in plant science in 2007 and his second degree from Jimaa university Natural Resources Management specialization of watershed management in 2011. He is senior watershed management expert at Oromia Regional state Bureau of Agriculture.

Amare Bantider

Amare Bantider (PhD) is Associate Professor of Geography and Natural Resource Management. He is senior researcher in Water and Land Resource Centre of Addis Ababa University and also lecture at the Center for Food Security Studies, College of Development Studies, Addis Ababa University. He obtained his first and second degrees from the Department of Geography and Environmental Studies at Addis Ababa University (Ethiopia) in 1987 and 1996 respectively. He obtained his PhD degree from Bern University, Switzerland in 2007. He authored and co-authored papers and book-chapters and published in peer reviewed journals and books on thematic areas of land use and land cover changes, watershed management, climate change, soil and water conservation, resource governance and related fields.

Messay Mulugeta

Messay Mulugeta (PhD) is an associate Professor of Socioeconomic Development, Food Security and Environmental Studies at CoDS/AAU. He obtained his first degree in Geography & Environmental Studies in 1997, second degrees in Environmental Science specialization in Environmental Resources Management in 2002, and his PhD degree in GeES (specialization in Socioeconomic Development Planning and Environment) in 2012 from Addis Ababa University. He is Ph.D Programme and Research Coordinator, Center for Food Security Studies, College of Development Studies, Addis Ababa University.

Ermias Teferi

Ermias Teferi (PhD) is a Director for Water Division in Water and Land Resource Center. He is also an Associate Professor at the Centre for Environment and Development, Addis Ababa University. He received his PhD in Hydrology and Water Resources from Technical University of Delft and UNESCO-IHE Institute for Water Education, the Netherlands, in July 2015. He was one of the recipients of the Netherlands Fellowship Programme (NFP) PhD fellowship award (2011) and WOTRO research grant award (2009).