Soil acidity under multiple land-uses: assessment of perceived causes and indicators, and nutrient dynamics in small-holders’ mixed-farming system of northwest Ethiopia

ABSTRACT

Knowledge on farmers’ perspectives is of paramount importance in order to design appropriate agricultural research and development interventions attuned to local farming systems. A participatory rural appraisal was conducted in order to understand perceived causes and indicators of soil acidity under multiple land-uses in three Districts of northwest Ethiopia. Soil samples were also collected from five dominant land-uses. The samples were analysed for soil pH, exchangeable acidity and other physico-chemical properties. The result indicated that the pH_(H2O) of most of the soils in the study sites were in a strongly acidic range (4.6–5.5). Gashena Akayita of the Banja District was the most acidic of all. Among the land-uses, eucalyptus fields were the most acidic followed by crop outfields and grazing lands in that order. At all the study sites, exchangeable Al was detected in soils having a pH of less than 5.0. Overall, the nutrient dynamics showed variation across land-uses and study sites. Farmers’ perceived causes of soil acidity included: soil erosion; contending use of fertility replenishing local resources; abandoning traditional fertility management practices and minimal use of external inputs. The farmers attributed the exclusive use of acid-forming inorganic fertilizers to exhaustion of the soil. Various land and soil characteristics, plant growth attributes, changes in genetic diversity were mentioned as indicators of soil acidity. Particularly, the farmers used prevalence of acidophilic weed species on crop fields and grazing lands as marker of strongly acidic soil. Farmers’ perceived causes and indicators were in agreement with scientific facts and can be utilized as input in designing sustainable acid soil management strategies. Decline in genetic diversity of the once widespread crop species and land races, and expansion of newly introduced soil acidity tolerant species, suggests the need to undertake rescue collections in these areas.

KEYWORDS:

• Ethiopia
• farmers’ perception
• participatory rural appraisal
• soil acidity

Introduction

Chemical degradation of land due to soil acidity is one aspect of land degradation constraining crop production worldwide. Acid soils constitute 30% of the world’s total ice-free land. In Africa, 22% or 659 million ha of land has a soil acidity problem (von Uexk¨ull & Mutert 1995). In Ethiopia, acid soils with a pH of below 5.5 in the surface layer constitute about 13.2% of the total land (Schlede 1989; Abebe 2007). In the country, acid soil-affected areas are found in the high rainfall parts, that is, west, northwest, southwest and south, which have good agricultural potential to offset poor productivity of areas under recurrent drought (Abebe 2007; IFPRI 2010). In addition to soil acidity, agriculture in these areas is constrained by land degradation associated to overgrazing, deforestation, continuous cultivation and limited or no use of external inputs that have resulted in serious problems of soil erosion and depletion of soil nutrients (Lakew et al. 2000; Girma 2001; Alemneh 2003). The ever-increasing demand for more agricultural land and degradation of existing ones suggest the need to address the problem through development of sustainable technologies and creation of favourable enabling environment (Nabhan 1999; Bruinsma 2009).

The farming system of acid soil-affected areas is exclusively mixed farming operated by small-scale farmers (Teklu et al. 2009; IFPRI 2010). Application of lime, mineral fertilizers and compost along with soil and water conservation measures were interventions under use to mitigate soil acidity in these areas (ATA 2013). Nonetheless, these interventions were mainly focused on the reclamation of crop fields. Earlier Alemneh (2003) also criticized that soil fertility management strategies of the government of Ethiopia gave no or little consideration to variations in agro ecologies, socio-economic situations and local resource endowments. On-going and emerging research activities are also basically focused on lime use (Achalu et al. 2013; Workneh 2013).

Farmers have indigenous knowledge and experiences on how to adapt to varied socio-economic and biophysical circumstances (Freudenberger 1994; Dixon et al. 2001). Hence, farmers’ participation in the process of problem identification helps to generate technologies that are compatible to a specific socio-economic and biophysical environment (Freudenberger 1994; Nabhan 1999; Dixon et al. 2001). Various studies also indicate correlation between farmers’ knowledge and scientific facts on causes and indicators of land degradation related to soil fertility decline (Malley et al. 2006; Karltun et al. 2011; Vaidya & Mayer 2014).

Land-use is one of the anthropogenic factors that affect soil reaction and other physico-chemical properties (Behera & Shukla 2015). In Ethiopia, research endeavours specifically targeting acid soil environments are rare and constrained by a paucity of information across different land-uses. This research was carried out in order to determine the state of soil acidity across multiple land-uses and to assess farmers’ knowledge and understanding on the soil acidity.

Material and methods

Description of the study sites

The study was conducted in areas affected by acid soils in the West Gojjam, Awi and East Gojjam administrative zones of the Amhara region in north-western Ethiopia (Figure 1). Based on available secondary data, and in consultation with experts from Agricultural Development Offices, one District was chosen from each administrative zone, that is, Mecha (West Gojjam), Banja (Awi) and Gozamin (East Gojjam). Finally, Enguti, Gashena Akayita and Enerata kebels ¹ were selected as study sites from Mecha, Banja and Gozamin Districts, respectively. The study was conducted in January 2013.

Figure 1. Geographical location and altitudinal range of the study areas (separate TIFF file provided).

According to the traditional agroecological classification which is mainly based on altitude the Dega and Woina Dega zones include areas with elevations of 2300–3200 m and 1500–2300 m above sea level, respectively (IFPRI & CSA 2006). Accordingly, Gashena Akayita (2500–2700 m) and Enerata (2400–2700 m) belonged to the Dega zone, and Enguti (1900–2100 m) was grouped in the Woinadega Zone. Based on temperature and moisture regime, they were classified into cool and sub-humid; cool and moist and tepid and moist agroecologies, respectively. Compared to the other two sites, Enguti allowed diverse crop choice due to its warmer temperature. The rainfall pattern is unimodal across all the study sites. Gashena Akayita received the highest rainfall (Figure 2). The mean land holding per household was 0.5 ha for Banja and 1.5 ha for Mecha and Gozamin Districts. The soil class of Gashena Akayita is predominantly Acrisol (Ultisol) whereas those of Enerata and Enguti are mostly Nitisols (Yihenew 2002; IFPRI & CSA 2006).

Figure 2. Rainfall pattern of the study areas: average of 10 years interpolated by FAO local climate estimator (FAO 2005).

Data collection

Data collection involved secondary data gathering from local sources on soil type, land-use, vegetation cover, major crops produced, animal and human population, etc. Primary data were collected through Participatory Rural Appraisal (PRA) approaches and soil sample collection and analysis as described below:

Direct observation

A transect-walk was made at each site with key informants to better understand the farming systems in terms of land-use, land form, vegetation cover, etc. through visual observation and discussion with people etc.

Group discussion

Group discussions involving about 20 farmers were held at each study site. Farmers’ perceived indicators, and causes of soil acidity, changing patterns in land-use, spatial distribution of soil acidity, etc. were assessed by using cause and effect analysis; trend analysis; listing and sorting PRA tools. The outcomes were established through discussion and consensus among participants.

Soil sample collection

Soil sampling strategy across the different land-uses was established based on secondary data analysis, and the outcome of the transact-walk and the group discussion. Accordingly, the land-use types were categorized into five classes: homestead plots, outfields, eucalyptus fields, natural forest and grazing land. Soil samples were collected from each land-use type except natural forest at Enerata due to lack of natural forest coverage in the area. A field or a plot with an area of about 2 ha was taken as a sampling unit (Motsara & Roy 2008). Each composite sample was constituted from a minimum of five fields for the crop and eucalyptus fields. For natural forest and grazing land, all available patches were sampled. From each sampling unit (2 ha area), a minimum of five samples were collected diagonally using a 20 cm auger. Overall, 14 composite samples were collected to represent the different land-uses and crop management systems across the 3 study areas.

Collected soil samples were air dried, crushed and sieved with a 2 mm diameter mesh and composited and thoroughly mixed for each land-use and study area. A quartering method was used to extract the final 1 kg soil needed for analysis.

Soil analysis

Soil samples were analysed for soil texture, pH, electrical conductivity (EC), exchangeable bases, exchangeable acidity, exchangeable Al, Cation exchange capacity (CEC), Organic carbon (OC) content, organic matter (OM) content, total nitrogen (TN), available phosphorous and micro nutrients (Zn, Mn, Cu, Fe) at the Amhara Design and Supervision Works Enterprise laboratory, Bahir Dar, Ethiopia.

FAO laboratory procedures outlined in Motsara and Roy (2008) were used to conduct the analysis of the specific physico-chemical properties. The pH was potentiometrically measured in the supernatant suspension of a 1:2.5 soil:H₂O/KCl mixture. Particle size was determined using the hydrometer method. CEC was determined by extracting with 1N ammonium acetate at pH 7.0 using a saturated titration method. Exchangeable bases in the ammonium acetate leachate were measured by an atomic absorption spectrophotometer (AAS). Micronutrients were extracted by DPTA (diethylene triamine penta acetic acid) and determined using AAS. Exchangeable acidity and exchangeable aluminium were determined by the titration method. Available P, OC and TN were determined using the Walkley and Black; Olsen and a semi-micro Kjeldahl methods, respectively.

Results

Predominant land-uses of the farming system

The proportion of land allocated to different agricultural uses ranged from 63% in Mecha to 93% in Banja (Table 1). Gozamin and Banja host zonal towns and consequently had a smaller rural people. Among agricultural land-uses, crop production stood first in terms of area coverage, followed by grazing land. However, the percentage of total land area allotted for crop production varied across the study sites. The largest share (46%) allotted for crop production was in Mecha and the least was in Gozamin (36.5%).

Table 1. Selected indicators of farming system and land-use characteristics of the study districts.

	Districts
	Gozamin	Banja	Mecha
Total Population^a	225,638	126,546	323,374
Rural Population (%)^a	63	77	91
Population density (persons/km^2)	182.0	249.1	218.3
Average land holding (ha)	<1.5	<0.5	<1.5
Total area for all land-uses (ha)	121,781(12.8)	30,217	156,027
Major agricultural land-uses (%)	76.4	93	63
Cultivated area (ha) (%)	44,488 (36.5)	12,277 (40.6)	72,138 (46.2)
Grazing land (ha) (%)	32,936 (27)	3443 (11.4)	15,591 (10)
Natural forest (ha) (%)		2679 (8.9)	5971 (3.8)
Plantation forest (ha) (%)	15,594(12.8)^b	9694.5 (32.1)	3106 (1.99)
Livestock population
Cattle	129,158	73,820	300,890
Shoats	121,998	102,416	175,619
Equines	21,838	26,622	35,355

Source: Respective zonal and district offices of agriculture.

^a2012 National Statistics (Abstract), CSA 2013.

^bTotal for both plantation and natural forest.

Crop production was principally cereal based. Potato (Solanum tuberosum L.) was the only non-cereal crop grown on substantial area of land across all the three study areas. Maize (Zea mays L.), finger millet [Eleusine coracana (L.) Gaertn] and tef (Eragrostis tef Zucc. Trotter) were the top three crops in terms of area coverage in Mecha district. At Gozamin wheat (Triticum aestvum L.), tef and maize were grown in large areas in that order. In Banja which is the most acid-affected area, tef, potato and triticale were widely grown. Brown-seeded tef land races were widely grown at Banja (Gashena Akayita) and Gozamin (Enerata) due to their adaptation to the edaphic and climatic conditions of these areas. Oat (Avena sativa L.) also covered a substantial area at Enerata. According to the farmers, adaptability to the agro-ecology was a key factor that affected crop selection and area allocation.

The grazing land is predominantly communal at all study areas. A larger percentage of total land area was allocated for grazing at Gozamin. Nonetheless, most of the communal grazing lands were overgrazed and severely eroded. Substantial proportions of grazing lands were occupied by weed species with no feed value. Collection of cattle dung dropped on grazing lands for fuel use was widely practiced. Due to the low productivity of grazing lands and high livestock populations, farmers relied on crop residue to feed their livestock. Most farmers at Banja used the most acidic outfields for pasture production as a coping strategy to soil acidity. Efforts to improve the productivity of communal and private grazing lands were also negligible.

Strong integration between crop production and livestock rearing was demonstrated in several ways at the study areas. Horses in the extreme highland parts and oxen in mid-altitude areas were used to provide draught power for land preparation. Threshing and transport of crops was also carried out by livestock. Equines were the predominant means to transport agricultural produce to the markets and important inputs such as fertilizers and lime to the farms.

Livestock also provided manure to replenish soil fertility, principally in homestead areas. In the Banja area, hura or night corralling of cattle on outfields was a longstanding popular practice used to replenish soil fertility through direct application of manure and urine.

The area covered by natural and planation forests in the Banja area was equivalent to the area allotted for crop production. In the other Districts, the area covered by natural forest was restricted to the surroundings of ancient churches and monasteries. The amount and even distribution of rainfall in this District could be the likely factor that contributed to the success of earlier afforestation programs. Banja was the most densely populated and had the smallest average land holding per household. The rainfall distribution that allowed double cropping might have lessened demand for more land. Potato was second to tef in the farming system of Banja. Potato as acid soil tolerant and a crop with high yield per unit area, could help sustain the local population for much of the year. Considerable rural households obtained income from the sale of bamboo- (Arundinaria spp.) and Eucalyptus-derived products. Plantations of green wattle (Acacia decurrence) were rapidly expanding on acidic soils that no longer support crop growth. Private eucalyptus plots constituted the prime plantation forest cover at Gozamin and Mecha. Soil acidity and prohibitive cost of fertilizer were among the main driving factors forcing farmers to switch to eucalyptus planting. Traditional agroforestry that involved deliberate maintenance of nitrogen fixing forest species such as Croton macrostychus and Cordia spp. etc. also contributed to the vegetation cover in the Mecha area.

Assessment of physico-chemical properties of the soils

The textural class of soils from Enerata and Enguti was clay whereas that of Akayita was mainly loamy. Variation was observed in the physical properties of soils across different land-uses. Outfields of Enerata belonged to the heavy clay class and had very low (<2%) OC, while homestead, grazing land and eucalyptus plots had a clay texture and a higher OC content. Generally, samples from homesteads and grazing land had relatively higher C contents (Table 2). Homestead soils benefit from organic matter applied in the form of household refuse and animal manure. Absence of tillage practice and poor decomposition of organic matter might contribute to, relatively, the higher C content of grazing lands in the study areas.

Table 2. Texture and organic matter and organic carbon proportion of soil samples across different sites and land-uses.

Sites	Land-uses	Texture (%)			Class	OC (%)	OMC (%)
		Sand	Clay	Silt
Enerata	Crop outfields	11	67	22	Heavy clay	0.82	1.41
	Homestead	19	53	28	Clay	2.22	3.83
	Grazing land	21	49	30	Clay	2.11	3.63
	Eucalyptus	14	59	27	Clay	1.95	3.36
Akayita	Crop outfields	13	21	66	Silt loam	3	5.18
	Homestead	61	11	28	Sandy loam	3.96	5.11
	Grazing land	35	21	44	Loam	3.61	6.22
	Eucalyptus	35	23	42	Loam	0.7	1.21
	N. forest	47	9	44	Loam	1.76	3.03
Enguti	Crop outfields	5	69	26	Heavy clay	2.57	4.44
	Homestead	19	53	28	Clay	2.94	5.08
	Grazing land	23	47	30	Clay	4.25	7.33
	Eucalyptus	9	65	26	Heavy clay	2.38	4.1
	N. forest	23	47	30	Clay	1.48	2.55

According to Landon (1991), carbon content of nearly all the study sites and land-uses fall in the very low (<2%) to low (2–4%) range (Table 2). Soils of Gashena Akayita had the highest content of organic carbon (OC) for all land-use types except for eucalyptus. The farmers reported heavy use crop residue for animal feed. Use of animal manure for fuel was also, reportedly, widespread. Hence, the overall low C content of the soils can be partly attributed to low return of organic matter to the soil. High frequency of tillage is commonly practiced for fine bed preparation required for the minute seeded and widely grown crop, tef. Such practice could also exacerbate loss of C through oxidation. Various studies have also indicated that increasing tillage intensity resulted in increasing rate of CO₂ emissions released from soil into the atmosphere (Liu et al. 2006; Krištof et al. 2014)

According to FAO classification outlined by Motsara and Roy (2008) the pH_(H2O) of most of the soils in the study sites was in the strongly acidic range (4.6–5.5) (Table 3). Only, one soil sample taken from homesteads at Enguti had a moderately acidic pH_(H2O) of 5.6–6.5. Sample from outfields of Gashena Akayita belonged to the extremely acid class (<4.6). Among the study sites, Gashena Akayita was found to be the most acidic environment, followed by Enguti and Enerata. Except soil from natural forest and homestead, the pH_(H2O)values of all samples from this site were below 5.0. Among land-uses, samples from outfields and eucalyptus plots were the most acidic (Table 3).

Table 3. Chemical properties of the soils across study sites and predominant land-uses.

Site	Land-use	PH (H2O) 1:2.5	PH (KCl) 1:2.5	ECe	Exchangeable bases (cmol(+)/Kg				Ca: mg	K: mg	% base Sat.	CEC cmol (+)/Kg	TN (%)	Av. P Olsen (mg kg^−1)	Ex. Al cmol (+)/Kg	Ex.Acicmol (+)/Kg	Micro nutrients (mg kg^−1)
					Ca	Mg	Na	K									Fe	Zn	Cu	Mn
Enerata	Crop outfields	5.28	3.89	0.17	14.71	2.59	0.33	0.47	5.68	0.18	55.86	32.4	0.04	4.98	0	1.2	15.4	0.6	1.3	12
	Homestead	5.49	4.17	0.68	16.54	2.87	0.03	2.11	5.76	0.74	74.31	29	0.1	24.62	0	0.64	27.8	1.4	2.3	21.6
	Grazing land	5.36	3.95	0.23	17.16	3.45	0.03	0.65	4.97	0.19	68.68	31	0.1	4.21	0	1.12	21.1	0.7	1.5	16.9
	Eucalyptus	4.74	3.42	0.45	11.52	2.69	0.19	0.52	4.28	0.19	51.45	29	0.08	4.4	1.92	4.24	18.9	0.5	1	13.8
G. Akayita	Crop outfields	4.57	3.73	1.24	1.56	6.93	0.08	0.47	0.23	0.07	21.02	43	0.27	18.82	1.6	3.84	28.3	0.7	0.7	2.6
	Homestead	5.16	4.12	1.68	12.16	0.76	0.07	1.52	16.0	2.00	38.18	38	0.25	34.6	1.76	4.4	39	1.5	1.6	7.9
	Grazing land	4.61	3.81	1.81	5.84	1.95	0	0.86	2.99	0.44	19.22	45	0.31	10.52	0.16	1.28	28.4	0.5	0.5	1.8
	Eucalyptus	4.82	3.81	0.86	6.09	6.64	0.09	0.35	0.92	0.05	30.63	43	0.06	13.67	1.36	3.36	27.6	0.4	0.4	1.5
	N. forest	5.44	4.62	1.43	33.74	4.75	0.08	0.95	7.10	0.20	73.19	54	0.76	3.18	0	0.48	35.9	1.6	1	7.1
Enguti	Crop outfields	4.9	3.68	0.29	7.24	1.56	0.05	0.52	4.64	0.33	30.62	30.6	0.11	3.98	0.56	2.24	12.6	1.5	1.9	21.5
	Homestead	5.86	4.67	0.68	17.58	3.23	0.24	2.1	5.44	0.65	74.68	31	0.13	11.23	0	0.48	22.7	2.9	3.5	32.5
	Grazing land	5.23	4.07	0.83	17.19	4.58	0.24	1.72	3.75	0.38	66.66	35.6	0.18	3.56	0	0.64	32.8	1.7	2.5	17.1
	Eucalyptus	4.98	3.79	0.29	8.61	1.98	0.17	1.18	4.35	0.60	42.04	28.4	0.1	2.92	0.24	1.36	17.8	1.1	2.3	22.2
	N. forest	5.5	4.54	0.75	21.23	6.22	0.08	0.81	3.41	0.13	79.61	35.6	0.13	1.37	0	0.4	28.1	1.4	3	28.9

At all the study sites, exchangeable Al was detected in soils having a pH of less than 5.0 (Table 3). At Enerata, only the soil samples collected from Eucalyptus fields were positive for exchangeable Al (1 cmol⁻¹ of soil). At Enguti, the levels of exchangeable Al were 0.56 and 0.24 cmol kg⁻¹ of soil for crop outfields and Eucalyptus plots, respectively. At Gashena Akayita exchangeable Al was detected under all the land-uses, except in samples from natural forest. Since the soil samples were composites, the detected Al levels were mean values with the likelihood of an Al-toxicity problem in outfields of all the study sites, and those of Gashena Akayita in particular (Table 3). Similarly, Mn toxicity could be a potential problem at Enguti and Enerata.

Interpretation of available phosphorous varies depending on the crop demand (Sanchez 2007). Using Olsen’s method, an available P content of less than 4 mg kg⁻¹ of soil is deficient, and above 8 mg kg⁻¹ of soil is adequate for cereals. For potato, an available P content of less than 11 mg kg⁻¹ of soil is deficient, and above 21 mg kg⁻¹ is adequate (Cooke, 1967). Accordingly, except for homestead soils, the Nitisols of Enguti and Eneratahas P content were less than, or close to, the deficiency threshold. The highest levels of available P were obtained from homestead samples for all the study sites. At Gashena Akayita the P content of all the land-uses except that of natural forest was above 8 mg kg⁻¹. High available P content was previously reported for soils of Banja District (Gashena Akayita) (Yihenew 2002).

For total N content determined by the Kjeldahl method, N content of less than 0.1% is very low; 0.1–0.2% is low; 0.2–0.5 medium; 0.5–1 high and above 1% is very high (Landon 1991). Accordingly, the N content of Enerata, Enguti and Gashena Akayitawas is predominantly in the very low, low and medium ranges, respectively. The overall assessment indicates that samples taken from homestead, natural forest and grazing land soils had relatively higher N levels (Table 3).

As per Landon’s (1991) classification, all the study sites and the land-uses had high (25–40 cmol kg⁻¹of soil) and very high (>40 cmol kg⁻¹of soil) CEC. Among the study sites, soils from Gashena Akayita had the least base saturation. Among all the land-uses, crop outfields and eucalyptus plots had the least base saturation. Even though the overall base saturation status is one of the indicators of soil fertility status, the relative balance among the bases is a more important indicator of soil fertility or availability of nutrients to plants (Landon 1991).

Exchangeable potassium percentage (EPP) (exchangeable potassium expressed as percentage of total CEC) of 2% was suggested to be a minimum level to avoid K deficiency in humid tropical soils (Boyer 1972). In this study, the EPP of the study sites and the different land-uses showed considerable variation. Crop outfields had EPP of less than 2% across all the study sites, implying the need to apply K on outfields. The highest EPP was recorded for homestead soils across all the sites.

Based on the critical limits of the DTPA-extractable micronutrients’ level described in Motsara and Roy (2008), the Mn contents of Enguti and Enerata sites were in the very high (>6 mg kg⁻¹) range. Only two samples from homestead and natural forest had an Mn content of >6 mg kg⁻¹ at Gashena Akayita. The rest of the samples from this site had Mn content in the medium range (1.2–3.5 mg kg⁻¹).

Causes of soil acidity as perceived by farmers

Soil erosion and unsustainable farming practices

Farmers stated loss of fertile top soil through runoff as one of the most common causes of soil acidity in the surveyed areas. According to the farmers, rugged terrain, high rainfall, cultivating steep slopes with limited conservation practices were the primary underlying factors of aggravated soil erosion. Land degradation on grazing lands was driven by lack of any sustainable grazing land management system and overgrazing associated to animal populations. As opposed to crop outfields, no soil conservation practice was observed on grazing lands across all the study sites.

Contending use of animal manure and crop residue

According to farmers, the rise in human population and pressing demand for food and fuel had resulted in extensive deforestation, decline in per capita land holding and shortage of fire-wood. Hence, the farmers were forced to use cattle dung and stalks of crops like maize for fuel than replenish the soil. Reportedly, resource-poor farmers collect dried cattle dung from communal grazing lands for sale on local markets. Since communal grazing lands were generally unproductive, farmers reported that they heavily rely on crop residue for animal feed. Thatching houses was also another use of for stalks of small cereals like wheat, barley and triticale that otherwise would be returned to the soil. Stalks of crops like wheat and triticale were preferred for roofing than feed. According to the CSA (2007), 54% of residential houses in the Amhara region were thatch roofed (Figure 3). Furthermore, almost all the residential houses in the rural areas and small towns were mud-walled, plastered with straw of small cereals such as tef and finger millet. The mud walls and the floors of these houses were also painted with cattle dung to make them good looking. Straws of tef, finger millet and other small cereals were also used for preparing mattresses used in rural villages and small towns (Supplementary online figures 1, 2 and 3).

Figure 3. Proportion of corrugated iron sheet roofed houses of the Amhara region (Source: CSA 2007).

Grain, pasture and crop residues generally have an alkaline pH due to their high content of basic minerals (Upiohn et al. 2005). Hence, decline in pH of soils of the study areas can be associated to the complete outflow of nutrients from the system in the form of grain and crop residue. And this association is authentic since the pH of the homestead soil samples across the study sites was higher than that of the other land-uses.

Abandoning traditional fertility management practices

According to the farmers, high population pressure and consequent shortage of arable land were the underlying factors to abandon traditional fertility management practices such as fallowing, ‘Chichet’ or ‘hura’ (night parking or corralling of cattle on out fields during the rainy season as a method of in situ manure application), and crop rotation, manure application. Variability in number of cattle holding per household and difficulty to reach a collective decision, and risk of theft and predator attack were among other causes that forced farmers to abandon these communal practice at Enerata and Enguti. In the Banja area, where soil acidity is an acute problem, this practice was maintained with no external support provided. Adequate amount and, relatively, even distribution of rainfall has enabled double cropping in the Banja area. Consequently, the farmers at Banja were willing to sacrifice part of the growing season to practice night parking or corralling (supplementary online figure 4). At Enerata and Enguti, coupled with the acute land shortage associated to high human population, the rainfall pattern did not allow double cropping and farmers were not willing to sacrifice the whole or part of the growing season for corralling.

According to the farmers, rotation of cereals with pulses and niger [Guizotia abyssinica (L. f.) Cass] is the traditional practice that had been primarily used to replenish soil fertility. However, the farmers stated that suitability of outfields for pulses declined in their area due to growing level of soil acidity. Consequently, production of legumes had been restricted to the homestead areas, and cropping cereal after cereal became a common practice on the outfields of the study areas. Even when the soil was not a problem, farmers were inclined towards the production of staple cereals needed for basic food security. And pulses were intercropped with crops like maize, potato and Brassica spp. on homestead for household consumption.

Limited use of external inputs

The farmers agreed that continuous and exploitative farming with little nutrient recycling characterized their crop production system. Di-ammonium phosphate (DAP) and urea accounted for 100% of the mineral fertilizer sold in the study areas. DAP provides P and N while urea supplies only N. Grains and biomass on the other hand remove basic cations such as K, Ca and Mg in addition to N, P and other nutrients (Upiohn et al. 2005). Furthermore, utilization of these sources (DAP and urea) was not optimal for all crops, land-uses and socio-economic groups. For instance, deliberate application of local or external inputs on communal grazing land is non-existent across all the study sites. Farmers in most of the study areas were reluctant to apply lime due to its high initial cost, and difficulties associated with its transport and application bulky volume needed to neutralize soil acidity. Outfields were generally more acidic and were often fertilized with mineral fertilizer compared to homestead soils that benefited more from manure and household refuse. Other reports have shown that that the low level of fertilizer usage is one of the causes of depletion of soil nutrients in Ethiopian agriculture (IFPRI 2010; ATA 2013).

Indicators of soil acidity as perceived by farmers

Land and soil characteristics

The farmers reported that bottomlands and flattop lands were considered to be more fertile and less acidic than sloped lands. According to the farmers, the bottomlands benefit from inflow of nutrient and organic matter through sedimentation from uplands whereas the flattop lands experience minimal outflow of nutrients through erosion.

Acidic outfields were described as ‘kelal’ or ‘forehe’ at Enguti. ‘Kelal’ refers to ease of ploughing and poor water-holding capacity. ‘Forehe’ refers to poor-quality soils lacking organic matter and microbial activity; possessing a high friability due to the absence of cohesive force to aggregate the soil together. Such soils were light red compared to reddish brown fertile outfields or deep brownish homestead soils. At Enerata such soils were called ‘borebore’. At Gashena Akayita, acid soils were described as ‘gibiz’ which literarily means ‘pretender’. Such soils physically tend to be fertile but in reality they were poor in crop response. Like the acid soils of Enerata and Enguti, gibiz soils were described as easy to plough and having poor water-holding capacity.

Plant growth and productivity attributes

Poor establishment, stunted growth, pale green young seedlings, poor stands and poor tillering capacity of cereals and, therefore, poor crop and straw yield were among major indicators of acid soils mentioned by the farmers.

Diminishing suitability of the soil for once popular crops such as barley, faba bean and field pea and the differential suitability of such soils for acid-tolerant crops such as triticale, oat and white lupin were also other indicators of soil acidity. At Enerata extremely acidic soils which hardly allow production of oat, triticale and lupin were described as ‘Yemote’ or ‘dead’, and such soils were often planted with eucalyptus. Farmers associated poor response or increasing demand of crops for mineral fertilizers with growing level of acidity.

Prevalence of specific weed species on crop fields was also another indicator of soil acidity. Farmers complained about prevalence of the weed couch grass (Cynodon dactylon) on the acidic outfields of Enerata and Enguti. At Gashena Akayita, corn spurry (Spergula arvensis L.), annual knawel (Scleranthus annuus L.) and tiny mousetail (Myosurus minimus) were the main indicators of outfield soils with high level of acidity. Poor growth and poor vegetation cover often with slippery algal growth, when wet, and invasion by weed species of no feed value were all indicators of soil acidity on grazing lands.

Decline in crop genetic diversity

According to the farmers, the introduction of new crop and forestry species to the farming systems was mainly related to soil acidity. Wofiye, Senefkolo, Limenish and Werenj were traditional barley cultivars that were widely grown and are currently non-existent. At Gashena Akayita, Sindemena, Temj, Saldini, Masno and Dubar traditional barley cultivars were widely grown. Among these, Saldmi and Temeje were disappearing due to the development of soil acidity. At Enerata, substantial areas of land covered by barley on the outfields have been replaced by oat. The local name of oat is Engido, which was derived from the Amharic word Engida, which means stranger or newcomer. Two oat cultivars, Chimburdi and Rejjimu engido, were grown in the Enerata areas. Triticale was also another introduced crop which was rapidly replacing the production of various indigenous crop species, as a result of its tolerance of soil acidity.

Farmers at the Enerata area identified Dabo, Tikur mure, Bursa or Sergegna, Zambi, Natchmure as landraces of tef grown in the area. However, Dabo was the most popular and widely grown land race. It was liked for its adaptation to the soil and early maturity. Nonetheless, this landrace was brown-seeded and fetches a lower price on the market as opposed to the white-seeded varieties. It also has short stature, provides little straw and is difficult to harvest. According to the farmers, there has been a decline in production of white-seeded and bursa or mixed-colour (Sergegna) landraces in their area due to their poor tolerance to acid soils. A similar pattern was also witnessed by farmers from Gashena Akayita. The effect of soil acidity on crop genetic resources diversity was pronounced in the highland areas, Enerata and Gashena Akayita, where the range of crop choices is limited due to low temperatures.

The rate of conversion of crop lands to eucalyptus and green wattle (Acacia decurrens) was alarming and was a threat to the remaining genetic resources in the wild and crop environment. Farmers reported substantial loss of genetic resources as a result of agricultural encroachment into the wild ecology and the selection pressure exerted by soil acidity. Contribution of natural forest to the overall vegetation of the study areas has been significantly reduced during the last half century and acid-tolerant forestry species such as eucalyptus has covered much of the study areas. Communal grazing lands have also been overgrazed, and invaded with plant species with no feed value that can thrive well on acidic soils and suppress the diversity of the natural pastures (supplementary online figure 5).

Discussion

The farming system of all the study areas can be identified as a highland temperate mixed farming system described by Dixon et al. (2001) (Figure 1; Table 1). By virtue of their suitability for human and animal health, the distribution of the human population had been generally skewed towards the highland areas (UNECA 1996; Pankhurt 2009). Human population density for all the Districts was higher than the national mean, 85 people per km², (http://country-facts.findthebest.com/l/84/Ethiopia) and mean land holding was below 2 ha (Table 1). Land degradation associated to soil erosion and nutrient depletion was reported to be a serious problem affecting this farming system (Schlede 1989; Bishaw 2001; Dixon et al. 2001; Paulos 2001; IFPRI 2010).

According to Landon (1991), carbon content of nearly all the study sites and land-uses was very low (Table 2). This can be partly attributed to low return of organic matter to the soil. High frequency of tillage practiced for tiny seeded and widely grown staple cereal, tef, could also exacerbate loss of C through oxidation and emission of CO₂ into the atmosphere (Liu et al. 2006; Krištof et al. 2014)

In this study, a soil test result indicated the prevalence of soil acidity across all the study areas and land-uses (Table 3). Despite their clay and heavy clay texture, the overall CEC of Enerata and Enguti sites was lower than that of Gashena Akayita. This can be partly attributed to relatively high organic matter content in soils from Gashena Akayita. Compared to organic matter, contribution of clay minerals to CEC is extremely low (Landon 1991). The higher exchangeable acidity at Gashena Akayita indicated greater contribution of acidic cations Al³⁺ and H⁺ to CEC of this site. Such soils need the application of high rate of lime to neutralize the high levels of Al³⁺ and H⁺ ions in the soil solution as well as on the exchange sites. But this high rate was prohibitive for small-scale farmers due to financial and technological limitations (Rao et al. 1993). Reluctance against lime application by most of the farmers noticed in the study areas could also be associated to these limitations. Exchangeable Al of 2–3 cmol kg⁻¹ of soil was considered to be excessive for some crop species (Chapman 1966). As the soil samples were composites, high exchangeable Al content are expected, particularly from the acidic Acrisols of Gashena Akayita. Crop outfields from Enerata and Enguti had pH_(H2O) levels of less than 5.5 and Mn content of >12 mg kg⁻¹(Table 3). Such combination of low pH and extremely high Mn content would result in Mn toxicity to most plants (Menzies 2003). When the pH falls below 5.0, Mn toxicity concurs with Al toxicity. But since tolerant plants can change the toxic divalent manganese to its unavailable form through increase of the rhizosphere pH, bulk soil data may not reflect the toxicity of Mn on roots (Menzies 2003).

The variability observed in the balance of the bases and micro-nutrients’ content across land-uses and study areas could be associated to intrinsic and extrinsic factors. For instance, the Nitisols of Enguti and Enerata showed similar patterns for available P and potentially toxic content of Mn, etc. However, Gashena Akayita, which represented the sub-humid climate and Acrisols of Enjibara was distinct. Unique and age-old management practices like corralling, widely practiced in the Gashena Akayita, might have contributed to the relatively better available P content. At Gashena Akayita, exchangeable Al³⁺ was detected from most of the land-uses compared to the other two sites. The contribution of intrinsic and anthropogenic activities to variations in important chemical properties was also reported for Indian acid soils (Behera & Shukla 2015). The result, generally, suggested poor availability or considerable imbalance of nutrients in outfields and eucalyptus plots.

Crop production on outfields generally seemed difficult without N application. But unless the current N sources changed, worsening of the acidity problem is imminent even for the less acidic fields. In addition to N deficiency and P fixation, imbalance and unavailability of Ca is also a likely problem for crop production on the outfields of the study areas. Hence, utilization of calcium ammonium nitrate 20% N and 6% Ca or calcium nitrate urea (calurea, 34% N, 10% Ca) along with other P sources may be recommendable N and Ca source for the study areas (Barak et al. 1997; Bolan & Hedley 2003).

Farmers and key informants identified soil erosion, and unsustainable farming practices; poor nutrient recycling, competing use of animal manure and crop residues; abandoning traditional fertility management practices; and limited use of external inputs as major causes of soil acidity. Reportedly, high rainfall, undulating land form, poor water-holding capacity of the soil and inadequate soil and water conservation practices have contributed to the severe loss of soil in the highlands of Ethiopia (Lakew et al. 2000; Bishaw 2001). Elsewhere, controlled experiments have indicated that runoff removes basic cations including liming materials and accelerates rate of acidity development (Anna et al. 1997; Ritchey et al. 2012). In Ethiopia, cultivated outfields are seriously affected by soil erosion with an estimated mean annual loss of 42 t.ha⁻¹ compared to 5 t.ha⁻¹ from pasture (Bishaw 2001). Hence, the high level of soil acidity on crop outfields revealed in this study can be partly explained by a high amount of rainfall and soil erosion (Figure 2).

Due to their distance from residential areas and difficulty to transport compost and manure, outfields were worst affected by poor recycling of basic cations in the form of organic matter (Teshome et al. 2014). Low organic matter content in soils of all the land-uses in the study areas also could have deteriorated soil physical properties and enhanced loss of basic cations through soil erosion and leaching.

Widespread utilizations of crop residue for animal feed and animal manure for fuel were the main contending uses of organic matter that otherwise would be used to replenish soil fertility (Supplementary online figures 1, 2, 3). Organic matter content of all dominant land-uses across all the study areas was low. Similar results were obtained in a previous study conducted in similar ecologies (Yihenew 2002). In the mixed farming system of the central part of Ethiopia, 70% of the total tef straw produced is used for animal feed (Zinash & Seyoum 1991). Grain, pasture and crop residues generally have an alkaline pH due to their high content of basic minerals (Upiohn et al. 2005). Hence, continuous removal of basic minerals in the form of grains and biomass subjects crop and grazing lands to increasing soil acidity (Murwira et al. 1993). Widespread use of cattle dung for fuel in the mixed farming system of Ethiopia has been reported in several studies (Schlede 1989; Bishaw 2001; Paulos 2001; IFPRI 2010). Zenebe (2007) estimated that the use of dung as fuel instead of fertilizer reduces the country’s agricultural GDP by 7%. The beneficial effects of manure for soil fertility is mainly related to its supply of P, basic cations such as Ca and Mg, organic matter and its contribution to the improvement of soil physical properties (Murwira et al. 1993; Giller et al. 1996). Increasing pressure on agricultural land has compelled farmers also to abandon traditional soil fertility replenishing practices such as fallowing, night-parking, crop rotation, etc. resulting in depletion of soil nutrients (Sanchez et al. 1997; Lakew et al. 2000).

Farmers did not mention leaching associated with high rainfall and the soil parent material as major causes of soil fertility decline or development of soil acidity. Soils in high rainfall and high temperature areas acidify faster because of high weathering and leaching of basic cations (Hede et al. 2001). The amount of rainfall in the study areas was generally high and less variable (Figure 2). Specifically the sub-humid agro-ecology of Banja (Gashena Akayita) receives over 2000 mm per annum whereas Mecha (Enguti) and Gozamin (Enerata) receive above 1200 m (IFPRI & CSA 2006). Higher level of soil acidity at Gashena Akayita compared to the other two study sites relates mainly to the higher rainfall (Figure 2).

Leaching of basic cations is high in clay minerals dominated by a 1:1 silicate layer such as Kaolinite which do not fix significant amount of basic cations compared to montmorillonitic or other 2:1 group clay particles (von Uexkull 1986). The soils of Enguti and Enerata were predominantly Nitisols while that of Gashena Akayitawas an Acrisol (Yihenew 2002; IFPRI & CSA 2006). Inherently, the clay assemblage of these two soil classes is dominated by Kaolinite (Driessen et al. 2001). Hence, these soils lose basic cations rapidly through leaching and hence acidify faster than soils with less drainage.

Neither farmers nor key informants directly implicated mineral fertilizers for the development of soil acidity. However, the farmers said that ‘mineral fertilizer is addictive and has already spoiled our soil’. This can be related to increasing level of acidity attributed to the acidifying effect of DAP and Urea which are generally grouped under acid-forming fertilizers. Assimilation of these fertilizers results in the net release of H⁺ into the rhizosphere increasing soil acidity (Marschner 1995; Barak et al. 1997; Bolan & Hedley 2003). The fact that the outfields had lower pH compared to other land-uses can also partly be ascribed to the acidifying effect of these fertilizers. Generally the result suggests contribution of both intrinsic and anthropogenic factors for the variability across study areas and land-uses.

Farmers used various terms that indicated physico-chemical and plant attributes to identify acid soils. As loss of basic cations through erosion is enhanced with increase in slope, the identification of slope as probable indicator of soil acidity by the farmers was valid. Abebe (2007) also indicated that acid soils were distributed on gentle to steep slopes parts of west, northwest, southwest and south parts of Ethiopia. As manure and compost utilization is labour demanding, they are often applied on gardens in mixed farming systems of small-scale farmers in Africa (Giller et al. 1996; Sanginga & Woomer 2009). Teshome et al. (2014) also reported similar practice in northwest Ethiopia. Consequently, garden or homestead soils were presumed to be in suitable pH range for crop growth in the study areas. Nonetheless, at the most acidic environment, Gashena Akayita, pH of the homestead soil was below 5.5 with high levels of exchangeable Al, suggesting the need to reconsider the current assumption, and formulate suitable management options. The various indicators of acid soils identified in this study, poor crop stands, stunted growth, reduced microbial activity, friability when ploughing, poor water-holding capacity (specifically of the Acrisols of Gashena Akayita), were scientifically valid (Little 1989; Driessen et al. 2001; Upiohn et al. 2005).Corn spurry (Spergula arvensis L.) and annual knawel (Scleranthus annuus L.) are acidophilic weed species that constrain crop production in acid soil areas (Čiuberkis 2001; Čiuberkis & Koncius 2006). Strong correlation of farmers’ indicator plants of soil fertility with soil analysis result was also reported (Karltun et al. 2011). Hence, utilization of these weeds as markers of acidic soils can be used as indicators of acid soils.

Changes in floristic composition of agricultural lands were mentioned as new developments that followed increasing degree of soil acidity. Such a trend can be associated with variable sensitivity of plant species to low pH and associated mineral toxicities (von Uexkull 1986; Rao et al. 1993; Upiohn et al. 2005). The declining role of legumes such as field pea and fab bean in the rotation system of the study areas could be partly associated to soil acidity and Al and Mn toxicities that affect adaption of both the host and the rhizobium strains (Hamdi 1982; von Uexkull 1986; Upiohn et al. 2005).

Soil acidity selects acidophilic plants that have the capacity to grow on acid soils due to their peculiar capacity to overcome Al-toxicity and low P availability (Houdijk et al. 1993; Roem & Berendse 2000). This selection pressure results in loss of calcicole species and variants in wild and cultivated areas. Reportedly, with no mention of underlying cause, oat has replaced a wide range of local crop species and landraces in the farming systems of the central highlands of Ethiopia (IBC 2007). Rapid expansion of triticale in the study areas has also been resulting in the loss of indigenous crop genetic resources.

Indirectly, development of soil acidity, and associated mineral toxicity and deficiency, can also force the farmers to encroach upon the remaining potential lands in the wild ecosystem and thereby decrease above-ground and below-ground biodiversity associated with the wild ecosystem (Sanchez 1995). Overgrazing can also be a selection force that can assist change in floristic composition of grazing lands to acid-tolerant species (Angassa 2014).

Switching strongly acidic crop outfields to eucalyptus plantations was one of the coping strategies of soil acidity as reported by the farmers. Eucalyptus is highly tolerant to acid soil and aluminium toxicity (Neves et al. 1982; Barros & Novais 1996). Leite et al. (2010) reported reductions in the exchangeable Ca²⁺, Mg²⁺ and K⁺ and increase in Al³⁺ and H⁺ contents as a consequence of Eucalyptus cultivation. The strongly acidic test results and high content of exchangeable Al in soil samples collected from eucalyptus in this study indicates the fact that eucalyptus could worsen soil acidity.

The trend of shifting to production of acid-tolerant crop species and landraces, and forestry species by farmers observed in this study (Supplementary on line figure 5) suggests the need to emulate the farmers’ indigenous coping strategy in developing improved acid-tolerant crop, forage and forestry species as components of sustainable acid soil management technologies across the different land-uses. Undertaking rescue collection mission to the study areas and similar areas is imperative in order to save crop genetic resources under threat.

Acknowledgments

We are grateful to Alliance for Green Revolution for Africa (AGRA) for financing this study as a component of a PhD study at the African Centre for Crop Improvement (ACCI), University of Kwa-Zulu Natal. We also thank Amhara Regional Agricultural Research Institute (ARARI) for logistical support. Special thanks go to the farmers at the respective study areas, Dr Minale Wondie, and other colleagues, individuals and institutions who directly or indirectly collaborated in this study.

Disclosure statement

No potential conflict of interest is expected by the authors.

Notes

1. Kebele is the smallest administrative unit in the government structure.