Edinburgh Research Explorer Productivity, nutritive value and economic potential of irrigated fodder in two regions of Ghana

An on-farm study was conducted in the northern and Upper East regions of Ghana to investigate the productivity and nutritive value of irrigated ruzi grass [ Urochloa ruz-iziensis (R. Germ. and C.M. Evrard) Crins] (syn. Brachiaria ruziziensis (R. Germ. and C.M. Evrard)] and forage sorghum ( Sorghum almum ) grasses as options against dry season feed scarcity and to understand associated market opportunities. Sixty participating farmers each established 100-m 2 plots which were sown at 15 kg ha –1 drilled in 60-cm rows in the dry season of 2016 and 2017. Irrigation was by flooding of soil surface every alternate day throughout the period of the trial. At both regions, herbage accumulation and nutritive value of forage species were

periods of crop failure, and help farmers to purchase inputs for improved crop yields (Adam et al., 2010;Karbo & Agyare, 2002).In the semi-arid regions of West Africa, especially in northern Ghana, feed is a major constraint to livestock production, particularly during the dry season due to the scarcity and decreasing quality of natural pasture (Ademosum, 1994).Natural pasture is characterized by seasonal, inter-and intraannual, and spatial variations, which are key limiting factors to its productivity (Hiermaux, 1996).This is associated with both the temporal and spatial distribution of rainfall (Bayala et al., 2014), resulting in an abundance of good quality material during the rainy season, and scarcity and poorquality material during the dry season.Crop residues, such as groundnut (Arachis hypogaea L.) and cowpea (Vigna unguiculata L.) haulms, pigeon pea (Cajanus cajan L.) straw, maize (Zea mays L.), sorghum (Sorghum bicolor L.) and millet (Pennisetum glaucum L. R. Br.) stovers, constitute the bulk of ruminant feed during the dry season in northern Ghana (Karbo & Agyare, 2002;Smith, 2010).Other available feed resources are agro-industrial by-products (AIBPs) from households and agro-processing industries.Besides the possible conflicting demands for use of cereal crop residue as livestock feed or as soil mulch, the challenge of collection, transportation, and conservation limit its efficient utilization in smallholder livestock production systems (Pathak et al., 2012) Irrigated farming as a supplement to rainfed production has been recommended to increase cropping intensity, and crop and fodder productivity.Pasture irrigation has been described as a promising technique that increases forage resource availability (Arya et al., 2011) and livestock productivity by maintaining forage crops during periods of drought, which is reflected in livestock production (Ferreira et al., 2020;Mochel Filho et al., 2016).
Irrigated agriculture is not new in most countries in Africa, especially in the West African subtropics, though it is usually associated with rice (Oryza sativa L.), maize, and vegetable production, and not forage production.Elsewhere, irrigation has been used extensively in dairy production (ABS, 2006).Previous work has shown that irrigation provided an increase of about 288% in forage production, corresponding to 3,000 kg ha −1 , in four cuts carried out in a period of about 4 mo in Brazil (Dantas et al., 2016).Work in Morroco indicates that irrigated forages can supply 40-50% of total livestock energy requirements, mainly from perennial alfalfa (Medicago sativa L.), berseem (Trifolium alexandrinum L.), and maize (Alaoui, 2009).In Ethiopia, small-scale irrigation supports intensification of the crop-livestock system in the highlands (Getnet et al., 2014;Hagos & Mamo, 2014).
Research and development work on irrigated forage production in West Africa is virtually nonexistent, even though the necessary components are all in place (dams, rivers, fodder species, etc.).There is very limited previous experience of irrigated fodder production, for example in Ghana, though a number of opportunities exist for its development.These include an emerging fodder market in the Upper East Region and a recent growing demand for feed for the increasing number of livestock in the peri-urban areas of Ghana, where livestock were not historically common and where conditions are less suitable for grazing.Therefore, an opportunity exists for spatial integration of fodder production into irrigated cropping systems in northern Ghana.
As climate variation is expected to become more frequent (Martin et al., 2016) impacting forage availability and quality, promotion of high-yielding forage species with improved

Core Ideas
• Productivity of ruzi grass and forage sorghum was evaluated under irrigation.
• Plant age at harvests affected regrowth ability of ruzi grass and forage sorghum.• Harvest at 8 wk allowed two 4-wk regrowth, best nutritive value and herbage accumulation.• Market potential for irrigated fodder varies spatially and is demand dependent.• Market prices of feeds and irrigated fodder is not always a reflection of the quality.
nutritive value that are also better adapted to drought and low fertility soils is a promising approach for improving animal productivity in arid and semi-arid zones.Urochloa is one of the important tropical forage grasses of African origin.It is widely cultivated in South America, Australia, and East Asia and has demonstrated success in transformation of beef and dairy industries (Ghimire et al. 2015).It produces large amounts of biomass and has a vigorous and deep root system, (Rosolem et al., 2017).It tends to be drought resistant and resilient in infertile soils and produces well with relatively low levels of fertilizer inputs and is sometimes referred to as a "climate smart" forage species (Ngila et al., 2016).Forage sorghum has been reported as another example of a pasture species with resistance to drought stress, which reduces the effects of the forage production seasonality and ensures the maintenance of animal production throughout the year (Lima et al., 2017).It is a short-lived perennial forage species, easily established from seed with rapid growth and fodder yield accumulation within the year of establishment.Forage sorghum is a prominent crop in the Brazilian agricultural sector for being a high energy grass, with high digestibility, productivity and adaptation to dry and warm environments (Lima et al., 2017).These two forage species were considered based on their ease of establishment through seeds, reported herbage yield and ability to thrive with minimal soil moisture in areas with limited annual rainfall such as the northern area of Ghana.
Harvesting at different weeks after planting (WAP) has been reported to impact both quantitative and qualitative traits of forages (Buxton, 1996;Buxton & O'Kiely, 2003).Many authors had reported different suitable WAPs when yield and quality were most stable for harvesting forage grasses especially Urochloa and forage sorghum in different parts of the world and under different seasons and fertilizer regimes (Perazzo et al., 2017;Ramírez de la Ribera et al., 2008;Zemene et al., 2020).Beyond the first harvest, ability to regrow under different conditions is considered desirable (Afzal et al., 2012;Foloni et al., 2008).For semiarid regions, where the rainy season is short (3-4 mo) regrowth capacity with high productivity, will reduce production costs and maximize the production system (Ahmeda et al., 2016;Rao et al. 1996).Botelho et al. (2010) had demonstrated such potential in Forage sorghum (Sorghum almum) in Brazil.Respati et al. (2018), observed similar potential of Urochloa in Indonesia.Repetitive defoliation of forage regrowth has been reported to have either negative or positive effects on the yield due to several factors such as the availability of nutrients in the soil, meristem, the length of growing season, the frequency and intensity of defoliation (Ferraro & Oesterheld, 2002).The current study builds on previous work and explores the effects on timing of harvest on irrigated Urochloa and forage sorghum in northern Ghana to assess possible impact on regrowth yield and other phenotypic traits, which might consequently modify the nutritional value.
The objective of study was to assess the potential of irrigated fodder in the West Africa subtropics through a case study in the northern and Upper East regions of Ghana, by quantifying the productivity and nutritive value of ruzi grass [Urochloa ruziziensis (R. Germ.& C.M. Evrard) Crins] and forage sorghum under irrigation.A further objective was to evaluate the economic potential of irrigated forages in the region by exploring the market opportunities and profitability of irrigated fodder as well as farmer motivation for production via market responses.

Study area
Northern Ghana is located in the Guinea Savannah ecological zone of Ghana which is largely characterized by lowland and grassland areas.The area has a relatively dry climate characterized by a prolonged dry season between November and March/April and a single rainy season that begins in May and ends in October, with an annual rainfall that ranges between 750 and 1,050 mm (Ghana Statistical Service, 2013).The monthly mean temperature ranges between 22 and 34 ˚C (Acheampong et al., 2014).The two specific communities where the study was conducted were Bihinayili in Savelugu district in the northern region and Zanlerigu in Nabdam district in Upper East region.Bihinayili is located about 10 km North of Tamale, the regional capital (9˚36′06.2"N0˚51′44.4″W)and is 162 m above sea level (masl).Zanlerigu is located about 15 km Northeast of Bolgatanga, the regional capital for the Upper East (10˚48 22.3′N and 0˚44 10 .9′W)and is 206 masl.
The vegetation consists mainly of grassland, especially savanna, with clusters of drought-resistant trees.The temperatures vary between 14 ˚C at night and 40 ˚C during the day.Much of the production of crops takes place in small-scale and rainfed systems (Månsson, 2011).
Both communities are located in the Guinea Savannah agro-ecological zone, Zanlerigu (Upper East) is relatively closer to the Sudan and Sahel agro-ecological zones than Bihinaayili (northern region), and the former receives less rainfall than the latter (Balana, et al., 2020).
The study sites were characterized by mixed crop-livestock production systems.The livestock reared are goats, sheep, pigs, donkeys, cattle, and poultry (Adzitey, 2013).The major crops cultivated are millet [Cenchrus americanus (L.) Morrone], maize (Zea mays L.), sorghum [Sorghum bicolor (L.) Moench], bean (Phaseolus vulgaris L.), tomato (Solanum lycopersicum L.), and vegetables, grown during the rainy season from May to September.The weather patterns restrict rainfed cropping to a single cropping season; therefore, irrigation is practiced during the off-season.Tomato, onion (Allium cepa L.), okra [Abelmoschus esculentus (L.) Moench], and other vegetables are occasionally planted during the off season for consumption and income with supplemental irrigation (Blümmel et al., 2018).Vegetable plots are irrigated by watering-can and surface irrigation methods and water lifted using buckets and motorized pumps from the shallow wells.
Irrigated land areas in both communities are in a particular low-land area of the communities which in total are <2 ha.These irrigated land sites in both communities are usually waterlogged and therefore unused during the rainy season.Each household with access to the irrigated land area map out individual plots and cultivate according to local agreement.Access to the limited irrigated area is by inheritance, or allocation through the local chiefs who act as the custodians of community land.Other criteria may be the ability to own small reservoirs or availability of vacant plots (Acheampong et al., 2018).

Land preparation and plot establishment
The study involved on-farm agronomic trials of two forage species under surface irrigation carried out on farmers' fields.The selected forage species were ruzi grass known for drought tolerance, adaptability to poor quality soil, and soil conservation properties; and forage sorghum, also known for its drought tolerance.Sixty farmers (30 from each region) who participated in the Feed the Future Innovation Lab on Small Scale Irrigation (ILSSI) project were selected based on willingness and readiness to allocate at least 100 m 2 of their irrigated land for the fodder trials for 2 yr.A total of 0.3 ha from the entire irrigated land area was used in each community.The field trials were carried out in October-April (dry season) for two consecutive years, in 2016 and 2017.Land preparation, field layout, plot irrigation, and sowing were performed manually in both regions.Nine representative soil samples were taken at 0-to-15-cm depth, three each from the top, middle, and lower slope of the experimental site in each region before the commencement of the trial, using soil auger.Composite soil samples from each slope per region were air-dried and passed through a 2-mm sieve before being analyzed.In addition to this, meteorological data of rainfall and temperature for the years of the field trials were provided upon request from the Ghana central meteorological station.

Soil physico-chemical characteristics
The physico-chemical characteristics of composite soil samples from the experimental sites showed that average total N recorded from northern and Upper East regions of Ghana was 200 and 100 mg kg −1 ), respectively (Table 1).The soil analysis in the northern region showed that soil contained 0.76 cmol kg −1 of exchangeable acidity, total exchangeable base of 3.93 cmol kg −1 and cation exchangeable capacity of 4.69 cmol kg −1 .The proportions of 0.54, 5.59, and 6.13 cmol kg −1 were respectively recorded at the Upper East region for exchangeable acidity, total exchangeable base, and cation exchangeable capacity.The rainfall distributions are presented in Figures 1 and 2. As presented, there was no rainfall in January and December for the 2 yr of measurement.Rainfall started in March and increased gradually to reach its peak between July and September 2015, after which it started to diminish.The rainfall pattern was largely unimodal.It should be noted that 3 of the 12 mo were completely dry for 2016 and 2017.The region had moisture to sustain vegetation growth throughout the year, but probably not sufficient for all year-round dryland forage production.Total rainfall in the northern region (1,084 and 919 mm in 2016 and 2017, respectively) were lower than that recorded (1,176 and 926 mm in 2016 and 2017, respectively) in the Upper East region.

Experimental design
The experiment was a factorial arrangement in a randomized complete block design (RCBD), comprised of the two regions: northern and Upper East regions of Ghana; two forage species: ruzi grass and forage sorghum; and four stages of harvest: 4, 8, 12, and 16 WAP.In each region, 30 plots were randomly selected for the trial with a total land area of 100 m 2 each, totaling 60 plots (30 per region).Each plot was marked out, cleared, and tilled manually in October 2016.Average cost of land preparation was determined for each region.Both ruzi grass and forage sorghum were established from seed by sowing at 15 kg per hectare, each on half of the plots.Ruzi grass was planted by row drilling at 60-cm spacing while forage sorghum was planted at 30-cm intervals in rows at 50-cm spacing.Each farmer's plot represented a replicate to account for the effect of land variation within the irrigated area.Acquisition of irrigation materials and equipment (pumps set up with inlet and outlet hoses) were arranged for both sites by the project.Cost of seeds, fertilizer, and fueling of water pumps were documented for each region.

Plot management
Weeding was carried out manually 3-4 WAP and as the need arose, using hoes and handpicks.The plots were fenced using locally available materials (Gliricidia poles and chainlink wires) in both districts.The cost of fencing was estimated and included in the total production costs.A compound fertilizer was applied to the forage grass plots at a rate of 20 kg N, 20 kg P, and 20 kg K ha −1 2 WAP.Irrigation was performed by flooding the soil surface following farmers' practice, using a 5-cm hose for 20-30 min every other day, using a shallow well as a water source.As the dry season progressed, hose size was increased to 10 cm.In each year, experimental design, seed rate, and plot allocation were the same and re-sowing was repeated only where necessary.

Agronomic data collection
Agronomic data including plant height, number of leaves per plant, leaf area, and number of tillers were collected in 2016 and 2017 seasons, in both regions at 4 and 8 WAP until when it was impossible to measure, especially Urochloa.Plant height of each forage species from the ground surface to apical meristem was measured using a meter rule graduated in centimeters.Leaf area was calculated manually using a nondestructive method by measuring leaf length x width x 0.75 (Oosterom et al., 2001).

Herbage accumulation
Herbage accumulation of each forage species was determined at four harvesting stages: 4, 8, 12, and 16 WAP.For each harvesting stage, regrowth at 4-wk intervals was additionally harvested until no substantial biomass from the plot could be recorded.Dry matter accumulation (DMA) at each harvesting stage and regrowth period were determined separately by harvesting the fresh forage 15 cm above the ground level within a 1-m 2 quadrat using a hand sickle.The fresh forage was weighed, and a random subsample of 300 g oven dried at 65 ˚C for 48 h reweighed and used to estimate DMA by multiplying fresh forage biomass by the respective DM concentration of the samples.

Laboratory analysis
Nutrient analysis of herbage was conducted at the nutrition laboratory of the International Livestock Research Institute (ILRI), Hyderabad, India.Subsamples from each harvested sample were dried at 65 ˚C for 72 h and then ground to pass through 1-mm sieve for the determination of N, ash, neutral detergent fiber (NDF) concentrations, in vitro digestible organic matter (IVDOM), and metabolizable energy (ME) using near infra-red spectroscopy (NIRS) equations developed for both grasses from conventional analysis of proximate chemical fractions (AOAC, 1990;Van Soest et al., 1994).The NIRS instrument used was FOSS Forage Analyzer 2500 with software package WinISI II.

Irrigated fodder market price
After the 16th weeks of the field trial (March 2017), forages were harvested from a 10-m 2 area, randomly selected from within the grass plots and from three different farmers' plots each, for both species, in both regions, for each year.
Harvested material was weighed and bundled for the market to assess buyers' estimated market price for each forage species daily for 5 d.At the market, the forages were rebundled according to the weight of fresh natural pasture available at the livestock feed market in Tamale, in the northern region and Bolgatanga in the Upper East region.Each bundle of irrigated forage was sold by the feed seller on the same day and the selling price per weight was recorded.Feeds available at the market during January-April 2017and their market prices were inventorized and categorized using a structured survey tool through personal interview with feed sellers.The sellers and buyers were also interviewed at the markets to assess their motivation to sell or buy irrigated fodder.Available feed samples sold at the market were bought from three to four different sellers.The price per kilogram DM of the purchased feeds was determined using at least three samples of each feed.Subsamples were later weighed and then air dried for nutrient analysis.Irrigated fodder farmers were also interviewed to assess their "willingness to continue".

Statistical analysis
Analysis of variance and Duncan's Multiple Range Test for mean separation using the general linear model of Statistical Analysis System (SAS, 2009) were used to analyze the phenological, DMA, and nutritive value variables.Each species was analyzed separately.Alpha level .05 was used where means are separated.The model for the statistical analysis included fixed variables (Region and WAP) and dependent variables.Significant interaction was determined using R statistical software (V 4.1).

Experimental model
Descriptive statistics were employed to present the qualitative variables obtained from the household survey.Ranking in the order of importance was done using normalized ranking.

Phenological characteristics
The mean plant heights, number of leaves and tillers, and the leaf area in both years increased from the 4 WAP to the 8 WAP (Table 2).At the 4 WAP, all measured phenological parameters, except the number of leaves in ruzi grass, were greater in the northern region than in the Upper East region.Conversely, the number of tillers of forage sorghum recorded in the Upper East region were greater than in the northern region, both at 4 WAP (5.9) and 8 WAP (6.5) while leaf area in both 4 and 8 WAP was significantly greater in the northern region than in the Upper East.It was observed that the plant heights in both species was consistently higher in the northern region than in the Upper East.

Herbage accumulation
When harvest was initiated at 4 WAP, regrowth allowed two subsequent 4-wk harvests both for ruzi grass and forage sorghum, in both regions, for both years.The DMA at each regrowth was similar to the first harvest, indicating that early harvest could produce similar productivity for at least two subsequent harvests at 4 wk cutting interval (Table 3).Harvesting both grass species at 8 WAP could only produce one further regrowth, no substantial yield could be recorded following a second regrowth period, in both regions with a lower DMA than the first harvest.Similarly, when harvesting was delayed until 12 and 16 WAP, it resulted in one harvest time with no substantial regrowth.Total cumulative DMA was lowest when the repeated harvests were initiated at 4 WAP (Table 3).Generally, ruzi grass gave greater (P < .05)DMA in the northern region than in the Upper East region across all the treatments.Dry matter accumulation of ruzi grass and forage sorghum were significantly (P < .05)affected by age at harvest, region, and their interactions.

Nutritional value
Crude protein(CP), ME, and IVDOM at different harvest stages were relatively similar between northern and Upper  4 and 5).Furthermore, CP and NDF means for ruzi grass were greater in the northern region than the Upper East although the regional differences was not significant for ME and IVDOM.This regional difference was also evident in the results for forage sorghum, except in the NDF concentration.Generally, 4-wk regrowth, whether after 4 or 8 WAP, had higher quality than either the first harvests at 8, 12, or 16 WAP.

Market values of irrigated fodder
The cost of production for each grass species was similar except for the cost of the seeds.Market prices for both grass species were also similar in each of the two regions in April, though lower in the Upper East compared with the northern region, which anecdotally correlates with the differences in nutritive values.Cost per kilogram of DM was different resulting from currency per yield per unit area of land and cost of the seeds (Table 6).Irrigated forages were priced lower in the Upper East region than in the northern region.The estimated market price of the irrigated fodder by the buyers at the livestock market showed that it has similar market value per unit kilogram DM relative to other available feeds in the market and with better nutritive value than cereal straw (Table 7).
The predominant livestock feeds available in the regions were groundnut haulm, cowpea haulm, soybean haulm, maize bran, cassava peel, pigeon pea straw, and rice bran (Table 7).In both markets, cowpea haulm is priced higher among the legume crop residues.There were fewer feed trading activities in the Upper East fodder market during this study as indicated by the types of feed resources available as well as lower prices (Table 7).

Relationship between the price and nutritional parameters of livestock feed resources
A scatterplot was used to determine the relationship between price (GHC/kg -1 DM) and N, ME, and IVDOM concentration of available feed resources sold at the Upper East and northern markets in Ghana in 2016 and 2017 (Figures 3, 4, and 5).The results showed a positive though weak linear relationship between price and all the nutritional parameters examined.
The relationship between the price and the N concentration, and ME were not significant (P > .05),but a significant relationship was recorded between the price and the IVDOM

Factors for selling and buying irrigated fodder
According to the livestock feed sellers in both markets, the leading motivational factors for selling irrigated fodder from both regions were high demand, regular supply of irrigated fodder in the market, and a high likelihood of profit (Figure 6).According to the respondents, provision of storage facility and access to credit were mentioned as factors that could stimulate their interest in selling the fodder.From the buyers interviewed, evidence of good performance of the animals fed irrigated fodder, regular supply of irrigated fodder in the market, and capacity building in conservation techniques of fodder were the major factors to stimulate buying irrigated fodder (Figure 7).

Factors that will promote irrigated fodder production
More than 40% of the farmers who participated in the irrigated fodder production from the Upper East region confirmed that guaranteed fodder market and evidence of profitability will motivate them to practice irrigated fodder production.Other factors include availability of forage seed, access to credit or financial services (e.g., subsidized inputs and building technical capacity).Fewer farmers suggested resolving issues surrounding land tenure security and availability of household labor as additional factors in promoting the production of irrigated fodder (Table 8).In the northern region, more than 60% of respondents chose guaranteed market for irrigated fodder and evidence of profitability as their motivation to continue in the production of irrigated fodder.A wide range of other factors were considered important in promoting the production.

Agronomic phase
The establishment and growth of both ruzi grass and forage sorghum under irrigation revealed their potential in the two regions of Ghana.Differences in the plant heights of both crops could be attributed to several factors among which could be differences in soil profiles in both regions.Soil texture at the Upper East region is a sandy-loam and neutral soil, while greater organic C and N content in the soil of the northern region is expected to support better crop production (Musinguz et al., 2013).The productivity of forage sorghum at 12 WAP was greater than what was reported for the same growth period in northern Nigeria by Ishiaku et al. (2016), but similar to the result of Olanite et al. (2010).Balole and Legwaila (2006) reported that forage sorghum yield annually about 20 Mg green matter ha -1 , which is similar to the DMA recorded at 8 WAP from the present study.The total DMA of irrigated forage sorghum in the Upper East was similar to that reported for forage sorghum when irrigated in the late dry season in the Sudan Savannah of Nigeria (Muhammad et al., 2005).

Cutting management and its effects on fodder yield and nutritive value
The result of the cutting management showed that both grasses responded to cutting management, both in quality and quantity.Nevertheless, the accumulated herbage DMA increased from 8 WAP and plateaued in ruzi grass but declined in forage sorghum beyond 8 WAP with advancement in maturity.As plants grew taller, they had more tillers per plant and more leaves per plant, resulting from accumulation of biomass components.This corroborates the report of Spitaleri et al. (1995), that herbage accumulation of grass increases with maturity, though the nutritive value declines.The results show that early harvesting promotes multiple regrowth but lowers total DMA.Multiple harvests may still be beneficial to forage farmers who target fresh fodder sales in the dry season at regular fodder markets in both regions of Ghana.Lower DMA might not sustain commercial livestock production especially in zero-grazing system in land-constrained areas.Delaying harvest until 8 WAP which produced single 4wk regrowth that provided good yield of higher quality forage in both species may be desirable by livestock farmers.
The biomass productivity differences between the two sites can be partly explained by the growth habit and adaptation of each of the grass species.Generally, ruzi grass performed better in the northern region in both years of study than in the Upper East region.Rao et al. (1996) suggested that Brachiaria species can adapt to a wide range of soil types from lowfertility acid soils to high-fertility neutral soils.They perform much better on acid soils (pH range of 5.0-6.8)than other grasses as the soils in this study from the northern region are more acidic.It is of paramount importance that high-yielding and palatable grass species should be established in suitable eco-sites (Mohammad & Naqvi, 1987).
Similar to DMA, the nutrient concentrations of forage are important to consider in selection of fodder varieties for livestock production (Maleko et al., 2014).Greater CP of the grass before 8 WAP is expected.Harvested grasses at 4 and 8 WAP resulted in younger plant regrowth material, which led to the greater CP and lower fiber concentrations.Late harvesting has been reported to significantly reduce CP concentrations of herbage species (Daniel, 1996;Teshome et al., 1994).In the 12 WAP treatment, the CP concentration of both grasses in Upper East region in both years of the study was below 75 g kg −1 which is the minimum level of CP for optimal rumen function (Adugna & Said, 1994;Van Soest, 1994).The implication is that supplementary protein-rich feed sources are required for optimal production if grasses harvested after 12 WAP are to be used as the basal diet for livestock in the region.Irrigated ruzi grass and forage sorghum harvested at an early growth/regrowth stage could be an alternative to continuous purchase of concentrate feeds by poor-resourced livestock farmers in the northern part of Ghana.
The NDF values fell within the range of 450-650 g kg −1 which is regarded as roughage feed of moderate quality (Turano et al., 2016).The trend of increasing NDF concentration with older harvesting age is a common phenomenon F I G U R E 6 perceived extent of of motivational factors for selling irrigated fodder by fed sellers in the northern and Upper East regions of Ghana (Amole et al., 2013;Asmare, 2016).The NDF concentration in tropical forages beyond which the DM intake of animals is negatively affected is considered to be around 650 g kg −1 (Van Soest et al., 1994).Higher NDF concentration occurs during later harvests, resulting in decreased DM digestibility (Mirza et al., 2002).The IVDOM concentration of the fodders was observed to decrease with age of the fodder.This corresponded with concomitant reductions in CP, ash, and ME and an increase in fiber proportions.Moore and Mott (1973) and Mugeriwa et al. (1978) reported that digestibil-ity values >650 g kg −1 indicate a better forage quality and that intake may be reduced below this.The values obtained at advanced harvest age (12 and 16 WAP) in the present study were by far lower than this critical level.It is important to note that decisions on the optimal time to harvest both ruzi grass and forage sorghum will depend on a compromise between productivity and nutritive value of the forage.Despite the reduction in CP concentration as harvest age increased, the CP yield increased as harvesting was delayed.The higher forage CP yield obtained for ruzi grass shows the potential protein benefits to be realized from allowing ruzi grass to grow until 16 WAP before it is harvested.This could possibly be a result of its late flowering as a perennial grass which kept the quality higher until this time.A relative reduction in the quality, especially CP concentration, once reproductive phase starts has been reported in most grasses (Amole et al., 2016).This phase is generally earlier in forage sorghum, consequently the CP yield declined earlier (12 WAP).Early reproductive development in forage sorghum leading to reduction in CP concentration, accumulation of cellulose and other complex carbohydrates, and lignification has been reported (Hazim & Ates, 2018).Asmare (2016) reported similar findings for desho grass (Pennisetum pedicellatum Trin.) when harvest was delayed until the 16th week after planting.
In any forage evaluation situation, compromises between quality and productivity must be made when deciding at what stage to harvest or graze a crop or pasture (Tilahun et al., 2017).In this study, harvesting at the early stage allowed multiple harvests, though lower total DMA.However, accumulated CP yield and ME yield were highest when both grasses were harvested twice (first harvest and one regrowth) and before the CP concentration and ME declined.In both regions this happened when ruzi grass and forage sorghum were harvested at 8 WAP, confirming that accumulation of CP is a function of total DM and CP concentration (Bayble et al., 2007).Though DMA and nutritive value are inversely related, technically evaluating CP and ME yields in relation to DMA provided the best harvesting time of the grasses.It therefore follows that farmers could use this information to assist in making decisions based on the relative importance of forage yield and quality in their operations.
The absence of cultivated fodder in the markets is an indication that commercial forage production is not routinely practiced in the study sites (Lukuyu et al., 2016) whereas there is a growing need for fodder especially during the dry season (Konlan et al., 2015).The scarcity of natural pastures has created a high demand for feed and has motivated feed sellers to collect browses, crop residues, and agro-industrial by-products (AIBPs) for small ruminant farmers (Husseini et al., 2011).Low feed trading activity and feed price in the Upper East region may be influenced by low demand, contrary to the report of Konlan et al. (2015).This study observed that demand and prices for fodder are higher in areas with limited access to free natural pasture due to urban development than in areas where livestock are raised mostly on free range (Tegegne et al., 2013).Differences in availability of feeds in fodder market and prices have been attributed to existing alternative feed resources in the same areas (Konlan et al., 2015).
An increment in the nutrient coefficients for IVDOM produced similar incremental increases in the price of the feed resources.Some have argued that there was no relationship between price and quality for all feed types in the fodder market survey peri-urban areas of Bamako in Mali (Ayantunde et al., 2014).However, this result is consistent with the report of Samireddypalle et al. (2017) suggesting that haulm fodder quality is the guiding principle for pricing in feed markets in northern Nigeria, especially with reference to digestible organic matter.Generally, pricing may be based on units of measure, physical appearance, and indigenous knowledge of acceptability by the animal and not actual quality analyses.Livestock keepers are not yet well informed on the nutritional quality of feeds (Ayantunde et al., 2014).Consequently, when feed prices are based on indigenous knowledge, quality fodder may be priced less than it could be, resulting in unequal pricing.The irrigated fodder harvested at 16 WAP for the market potential affect its pricing as older or stemmy material may be priced lower based indigenous knowledge.The linear relationship between price and quality in this study is an indication that any changes in the quality attributes may result in a corresponding change in the price.Creating opportunities for improving the quality of the feed resources in regions can lead to increases in the profit derived.
Attempts to market the harvested irrigated fodder demonstrated the potential demand for fresh fodder and its market acceptability.Regarding the market value of irrigated fodder, it was observed that this could serve as means of diversifying household income sources from the sale of fodder for producers having access to water source and available land.Yield per unit area and the market price at different location influence the cost benefit of engaging in irrigated forage production as demonstrated in this study.Grass species with greater yield per unit area, lower cost of production and higher return on investment, may be the irrigated farmers' preference especially in the northern region.
Although the market value of irrigated fodder is relatively higher than those of dry crop residues and bran, making it economically viable, the return on investment needs to be compared with other options with little or no cost of investment (roadside natural pasture, crop residues etc.).Similarly, where irrigated fodder production competes with more traditional irrigated commodities for land, water, labor, cash and other resources, farmers make decisions by considering their comparative advantages.With this situation in mind, capacity development in terms of awareness of the nutritional values of irrigated fodder, economic potential and technical support will be initial steps in promoting irrigated fodder trading in regions where it does not exist.Land size is one of the major limiting or enhancing factors.Especially in the Upper East region, available land areas for irrigated farming are located in the lowlands and are usually small as a result of fragmentation due to population growth.
Other limiting factors observed during the study were the farmers' livestock production targets in both regions which were not market oriented.In most studies where the profitability of irrigated fodder has been reported and is well established, livestock production is market oriented-mostly dairy production, even at the smallholder level (SNV, 2013).Others also report on the profitability of irrigated forage in eastern Africa (Getnet et al., 2014;Bezabih et al., 2016).Integrating fodder into irrigated production practices in both regions will require a change of mind-set of smallholder farmers into business-oriented livestock management.This is because for a successful adoption of irrigated fodder, farmers need sufficient land, water, and productive animals, and proper input and output market linkages (Gebrehaweria & Haileslassie, 2014).

CONCLUSION
As the demand for feed continues to grow with prolonged dry seaosn especially in the semi-arid regions, this trial showed greater potential for cultivation and marketing of irrigated fodder in places where such practices are uncommon.Ruzi grass and forage sorghum evaluated in this study showed great potential as feed resources in view of their high yields and quality especially under irrigation.Finding balance between DMA and nutritive values of forages requires targeting harvest at the optimum time.Harvesting at 8 WAP appear as the best time as it resulted into two harvests with reasonable quality materials.Market potential for irrigated fodder during the dry season varies spatially and is demand dependent.Feed pricing in this region is based on indigenous knowledge and not on determined nutritive value which may result in unequal pricing of the irrigated fodder.There is need for capacity development on informed nutritional quality of feeds and forages in region.This is important to accelerate the market value for irrigated fodder when harvested at the optimum time (8 WAP) in term of quality and quantity as evidence of good performance of the animals fed irrigated fodder could stimulate higher demand for irrigated fodder.However, where pricing is not a reflection of the forage quality, delaying harvesting till 16 WAP to generate higher DMA may be advisable.

A C K N O W L E D G M E N T S
This work received financial support from the U.S. Government's Feed the Future Innovation Lab for Small Scale Irrigation.

F I G U R E 1
Mean monthly rainfall distribution in millimeters for 2016-2017 in the northern region of Ghana F I G U R E 2 Mean monthly rainfall distribution in millimeters for 2016-2017 in the Upper East region of Ghana

F
I G U R E 3 Relationship between price and N concentration of feeds sold in the feed markets of the northern and Upper East regions of Ghana in 2016 and 2017 F I G U R E 4 Relationship between price and metabolizable energy (ME) of feeds sold in the feed markets of the northern and Upper East regions of Ghana in 2016 and 2017

F
I G U R E 5 Relationship between price and in vitro digestible organic matter (IVDOM) of feeds sold in the feed markets of the northern and Upper East regions of Ghana in 2016 and 2017 Amole: Conceptualization; Investigation; Project administration; Writing-original draft; Writing-review & editing; Emmaunel Panyan: Investigation; Supervision; Adetayo Adekeye: Data curation; Writing-review & editing; Augustine Ayantunde: Conceptualization; Methodology; Supervision; Validation; Writing-review & editing; Alan Duncan: Conceptualization; Methodology; Supervision; Validation; Writing-review & editing; Michael Blummel: Conceptualization; Methodology; Project administration.C O N F L I C T O F I N T E R E S T Authors declare no conflict of interest in relation to this publication.O R C I D T. Adegoke Amole https://orcid.org/0000-0002-0686-4913Alan Duncan https://orcid.org/0000-0002-3954-3067

T A B L E 1 Physico-chemical characteristics of the composite soil samples in study sites a Parameters Northern region Upper East region Method
a Source: Field data, 2016.

Ruzi grass Forage sorghum Parameters Age Northern region Upper East region Average P values SEM Northern region Upper East region Average P values SEM
Average phenological characteristics of ruzi grass and forage sorghum in the northern and Upper East regions of Ghana in 2016 and 2017Note.WAP, weeks after planting; SEM, standard error of mean.Average dry matter accumulation at each harvest (Mg DM ha −1 ) and total accumulation (Mg DM ha −1 ) of ruzi grass and forage sorghum in northern and Upper East regions of Ghana in 2016 and 2017 T A B L E 2 * Significant level based on P values at .05.T A B L E 3 *Note.WAP, weeks after planting; SEM, standard error of mean.* Significant at .05. ** Significant .01. *** Significant .001.

Cost of production Northern region Upper East region Ruzi grass Forage sorghum Ruzi grass Forage sorghum
Cost of production, sales, and economic benefit of irrigated fodder in the northern and Upper East regions in Ghana Cedi (GH¢ yield -1 unit area -1 ) T A B L E 6Note.DM, dry matter.a No cost of labor and fencing materials.b 1 GH¢ = US$0.22.

Crop residues Northern region Upper East region GH¢
Available feed resources and average price per unit of the feed in both regions kg −1 dry matter a T A B L E 7Note.na, not available at the site.a 1 Ghana Cedi (GH¢) = US$0.22.
Respondents' perceived extent of importance of motivational factors for buying irrigated fodder in northern and Upper East regions of Ghana Factors that will promote irrigated production of livestock feed by producer (N = 30 in each region)