Evaluating the Impact of Soil Fertility Heterogeneity on Maize Nutrient Requirement and Productivity in Smallholder Farming Systems

Most smallholder farms in Zimbabwe are on sandy soils which are inherently infertile, with soil organic carbon (often < 0.4 %) (Grant, 1981; Mtambanengwe and Mapfumo, 2005). The soils have poor buffering capacity, are subject to nutrient leaching, and have low cation exchange capacity. A large proportion of farmers are resource-constrained, with poor capacity to purchase external mineral fertilizers. Other than the macro nutrients (N, P, K), micro nutrient deficiencies are also widespread as a result of long-term nutrient mining through crop off-take without replacement (Grant, 1981).

In many smallholder farming areas across much of sub-Saharan Africa, marked short range spatial variability in soils exists within and between farms as a result of management and or localized differences in parent material (Giller et al., 2006; Wopereis et al., 2006; Zingore et al., 2007). Recognition of the variability in soils is important for implementation of management strategies that enhance efficient use of scarce nutrient resources on smallholder farms, taking into account the distinct capacity of the different soils to supply nutrients to crops (Janssen et al., 1990). When background soil characteristics or soil fertility indicators are known, it is possible to use them and appropriately tailor fertilizer application for improved nutrient use efficiency, thus ensuring that fertilizer interventions remain economically viable for smallholder farmers.

The project will explore nutrient management strategies for optimizing maize productivity under variable soil fertility conditions in eastern Zimbabwe. Multi-location nutrient omission trials will be conducted to diagnose nutrient and non-nutrient constraints to crop productivity. The nutrient omission trials will be used as a basis for establish maximum attainable yield for heterogenous fields and develop site-specific nutrient management recommendations for reducing the gap between attainable and farmers yields. Further, the project will develop and test decision support tools, including Nutrient Expert for Hybrid Maize, and extension material for supporting farmers to make informed decisions on investment in nutrient management. Spatial analysis will also be conducted to develop a village-scale soil fertility map and spatially explicit nutrient management recommendations.



Justification

While the majority of rural communities will continue to depend on agriculture to support livelihoods in the foreseeable future, the soil resource base has been depleted due to over 50 years of extractive farming practices, resulting in diminished crop productivity. Also, the wide variability in soil fertility on farms requires innovative strategies to enable efficient use fertilizer. Many studies have clearly documented the variability of soil fertility within and between farms in many parts of Africa (e.g. Tittonell et al., 2005; Mtambanengwe and Mapfumo, 2005; Zingore, 2006).

Spatial variability in soils on farms, largely due to differential nutrient management, has largely been ignored when designing technological interventions in smallholder farming systems. This is despite the major challenge for efficient use of fertilizer posed by such variability. In the face of limited availability of fertilizer, the question of efficient targeting to land units varying in soil fertility is critical (Giller et al., 2006). An approach towards mitigating such concerns is site specific nutrient management that encompasses more precise nutrient targeting as compared to the current practices that are largely based on N-P-K blanket fertilizer recommendations. A key step in development of site specific fertilizer requirements is the estimation of the indigenous nutrient supplies for specific fields, which is the cumulative amount of a nutrient from all indigenous sources (Dobermann et al., 2004; Sen and Majumdar, 2006). Application of fertilizer is then targeted based on nutrient requirement to achieve a target yield, set based on attainable yield and farmers economic and food requirements.

Glaring contrasts have repeatedly emerged between crop productivity on smallholder farms and that from experimental stations. Yields obtained in farmers’ fields are often <0.8 t ha^-1, compared with >5 t ha^-1 on research stations. The 4R Nutrient Stewardship Framework developed by the fertilizer industry provides a basis for efficient fertilizer use, focusing on four central management components: applying the right fertilizer source at the right rate, at the right time in the growing season, and in the right place. This project seeks to validate key components of the 4R nutrient stewardship and provide a basis for development of site-specific fertilizer recommendations in spatially variable smallholder farming systems. Experimental programs with fertilizers and other nutrient resources in Zimbabwe have largely focused on N, P, K and S with little or no reference to micronutrients or Ca and Mg yet these are essential to balanced nutrition. Yield improvement by up to 40% have been observed after deficiencies of B, Zn and S were removed through fertilizer application based on site specific soil sampling (Wendt, 1993).



Objectives

1. Conduct nutrient omission trials to diagnose nutrient and non-nutrient constraints to crop productivity.
2. Establish potential, attainable and farmers’ yields and develop nutrient management recommendations for reducing the gap between attainable and farmers yields.
3. Develop and test decision support tools, including Nutrient Expert for Hybrid Maize, and extension material for supporting farmers to make informed decisions on investment in nutrient management.
4. Conduct spatial analysis to develop a village-scale soil fertility map and spatially explicit nutrient management recommendations.

Methodology

Site description
The project will be part of Maize Coordinated Experiments by the IPNI Africa program to establish potential and attainable maize yields in the sub-humid zone in East and Southern Africa (Kenya, Malawi, Zimbabwe, Tanzania) and identify best nutrient management practices for optimizing maize productivity. The project will be carried out in Wedza (Goto ward) smallholder farming community in eastern Zimbabwe with about 750 mm rainfall annum^-1. Goto lies in natural region III with predominantly sandy Lixisols with challenges of low soil organic carbon and inherently poor nutrient supply. Soil organic carbon is typically between 0.3 and 0.5% on sandy soils with C> 0.5% on small pockets of loamy to clay soils. The study will be carried out within the framework of the Soil Fertility Consortium for Southern Africa (SOFECSA).

Site selection
In the first year, a soil survey will be carried out in a selected village by sampling soils from 64 fields within a 1 km² area. The fields will be selected in grid of 100 m x 100 m, excluding grazeland and land not used for cropping. Soil samples will be collected from a depth of 0-20cm and analyzed for of SOC. The SOC contents will be used to group fields into three categories as follows:

· Type 1: a field with < 0.46 C - In previous studies it was established that 0.46% C is the threshold below which response to fertilizers is inconsistent (Mtambanengwe and Mapfumo, 2005)
· Type 2: a field with 0.46-0.65 C – fields that have received organic amendments intermittently
· Type 3: a field with > 0.65 % C – these are fields that normally have a history of good management

Five fields representing each of the three SOC categories will be selected for establishment of experiments.

Experimental treatments
Nutrient omission experiment
A nutrient omission trial design with elements in Table 1 will be established on the 15 sites to determine indigenous supply potential for N, P and K and generate data for calibrating the Nutrient Expert for Maize DST. High rates of fertilizer will be used to enable determination of the maximum attainable yields for the various categories of soil fertility. The design of the experiment will be a randomized block design with the three replicates.

The SC513 maize variety with a yield potential of 8 t ha^-1 will be planted. Plot size will be 4.5 x 5 m². The experimental treatments will be replicated three times on each field. Maize crop spacing will be 0.75 x 0.25 m² to give a population of 53 000 plants ha¹. Target nutrient application rates will be 40kg P, 60 kg K₂O and 140 kg N. N fertilizer will be applied in three splits at proportions of 30:40:30 at 2 WAE, 5-6 WAE and at 8 WAE respectively.


Table 1: Treatments and fertilizer combinations

Treatments	Fertilizer materials and combinations
1. Control	None
2. NK	K2O
3. PK	TSP+ K2O fertilizer blend
4. NP	TSP + AN
5. NPK S Zn B Ca Mg	Compound D (NPK) + AN + ZnSO4 + CaSO4 MgSO4 + borax
6. Manure only	Manure at 5 t ha-1
7. Manure + NPKS Zn	Manure at 5 t ha-1 + Compound D (NPK) + AN + ZnSO4 + CaSO4 MgSO4 + borax
8. MgCaCO3 +NPKS Zn	Dolomitic lime + Compound D (NPK) + AN + ZnSO4 + CaSO4 MgSO4 + borax

Spatial soil fertility variability
Soil samples collected from the 1 km² areas in Goto ward will be geo-referenced using a GPS. Geo-statistics in Arc Gis will be used to assess spatial variability. Descriptive statistics together with interpolation of soil analysis data will be used to convert discrete sample data into continuous variability maps that allow demarcation of land units for nutrient management recommendations.

Total organic carbon
Organic C will be determined by the Walkley –Black method (Nelson and Sommers, 1975). A mixture of sulphuric acid and aqueous potassium dichromate (K₂Cr₂O) will be used to oxidize the carbon. After complete oxidation from the heat of solution and external heating, the unused K₂CrO₇ will be titrated against ferrous ammonium sulphate and a graph of absorbance will be plotted against a set of standards. The % organic C will be calculated as follows: % organic C = (K x 0.10)/ (W x 0.74) where K = sample concentration-mean blank concentration, W = weight of soil.

Soil characterization
Soil pH will be determined using the CaCl₂ method (Okalebo et al., 2002). A soil sample weighing 10g will be obtained and 25 ml of 0.01 M CaCl₂ will be added. The mixture will be shaken on a mechanical shaker for 30 minutes and thereafter a pH value will be obtained. The hydrometer method will be used to determine soil texture. A soil sample weighing 100g is to be weighed and 10ml of sodium hexametaphosphate added to the sample. The soil is to be saturated with distilled water and left to stand overnight. The mixture will be transferred into a container and made up to the mark with distilled water. The mixture will be mixed on a mechanical shaker for exactly 4_1/2 minutes and after mixing, a hydrometer will be inserted into the mixture. A temperature reading will be obtained after 30 seconds and thereafter the mixture will be left to stand for 2 hours and readings from the hydrometer and thermometer will be obtained. The percentage of silt and clay will be calculated and the textural class for the sand will be obtained from a textural triangle.

Colorimetric determination of total nitrogen and total phosphorous.
The content of total N and P is to be measured in a digest obtained by treating soil and plant samples with hydrogen peroxide, sulphuric acid and lithium sulphate. Selenium powder will be used as a catalyst. An air dried sample (0, 5g) mixture will be weighed into a digestion tube followed by addition of 4.4ml digestion mixture. The resultant mixture will be placed on a digester at 360⁰ C for 2 hours. After cooling the solution the mixture will be allowed to settle. Graphs of absorbance will be plotted against standard concentration.

Determination of total exchangeable bases
A soil sample will be extracted with an excess of 1M Ammonium Acetate. The amounts of exchangeable Na, K, Ca and Mg in the extract will be determined by flame photometry (Na and K) and atomic absorption spectrophotometry (Ca and Mg). Lanthanum and strontium will be added as a releasing agent to prevent formation of refractory compounds e.g. phosphates, which may interfere with the determination. The total exchangeable bases will be calculated as follows: mg per kg K, Na, Ca and Mg = {(a-b)*v * f * 1000}/ (1000 * w ) where a is the concentration of K, Na, Mg and Ca in the sample, b is the concentration of element in blank sample, v is volume of the extract solution, w is weight of the soil sample and f is the dilution factor.

Maize tissue and grain nutrient concentration
Atomic absorption, (emission for K) spectrophotometry through ashing will determine K, Ca, Mg and Zn. Nutrient concentration in maize tissue will be interpreted as deficient, low or adequate (Okalebo 2002).

Expected Outputs

• Attainable and target yields defined for fields that vary in soil fertility.
• Indigenous nutrient supply potential and nutrient sources and rates required for attunement of target yields developed under variable soil fertility conditions.
• Nutrient Expert for Hybrid Maize and extension material developed by the Africa program validated for East Zimbabwe.
• A village-scale framework for systematic development and presentation of site-specific nutrient management recommendation for smallholder farming systems developed.