1 ICRISAT, PO Box 776, Bulawayo, Zimbabwe; pres address 31 Woonalee St., Kenmore, Qld, Aust
Email rjkmyers@bigpond.net.au
2 ICRISAT, PO Box 776, Bulawayo, Zimbabwe Email g.heinrich@cgiar.org, j.rusike@cgiar.org
We report on farmer-participatory testing of strategies to increase manure use on crops to improve food sufficiency in communal farming lands of semi-arid southern Zimbabwe. In seeking to overcome farmers’ reluctance to use kraal manure, we tested cattle and goat manure, and methods of preparation (composted in heaps, heaps covered with soil, and in pits) on farmers’ fields in Tsholotsho and Gwanda South. Most methods increased yields, and farmers assessed them as practical and effective. Modelling with APSIM helped promote engagement with farmers. Despite drought and economic conditions, the combination of three research approaches – on-farm participatory trials, modelling, and farmer surveys – has resulted in adoption of manure application, and opportunities for further adoption. While male-headed households tended to adopt fertilizer use, female-headed households accepted FYM. To help farmers invest, extension agents need to escape from ideal recommendations, and offer soil fertility improving technologies.
Farmers in dry areas of Zimbabwe can overcome reluctance to invest in soil fertility, and increase yields, by use of under-utilized manure, though capacity to invest varies between differently-resourced households.
Key words
Sorghum, Simulation, Households, On-farm, Economics
In Zimbabwe, 55% of land is semi-arid, with 63% of the rural population. These lands are recommended for semi-extensive and extensive farming, and have poor and erratic rainfall averaging 450-650 mm in agroecological region IV and <450 mm in region V. The soils are mainly derived from gneiss, with some deep sands, and some soils derived from basalt. Most are low fertility, after cultivation and erosion. Despite efforts to transfer fertilizer to farmers, few apply it, as it is risky and gives little return under low rainfall. Few farmers have utilized manures (FYM) from kraals believing them to be risky. We sought solutions to low yields, food insecurity and poverty in communal areas. Here we examine (1) results of on-farm testing of FYM, and (2) using simulation to increase efficiency of on-farm studies. We used a participatory approach with emphasis on agronomy, socio-economics and technology transfer. FYM methods, chosen by farmers, were evaluated by researchers and farmers in trials on farmers’ fields.
We had two locations: Tsholotsho (lat 19 46 S, long 27 44 E, alt 1090 m) av 650 mm rainfall, with cambisols, luvisols, regosols and phaeozems (WRB), and Gwanda South (lat 21 34 S, long 29 02 E, alt 935 m) av 500 mm rainfall, with cambisols and luvisols (WRB). Table 1 is a weather summary in the 1999-00 and 2000-01. From 01 Dec to 31 Mar approximates the sorghum growing season. At Tsholotsho the 1999-00 season was dry, whereas 2000-01 was mildly droughty. At Gwanda South, the 1999-00 season was dry, and 2000-01 was droughty because of poor distribution. The technologies reported here included goat and cattle FYM, and 3 methods of FYM composting – heaped and covered, heaped and uncovered, deposited in a pit and covered. The trials were chosen by farmers who indicated a willingness to test manure.
Table 1. Weather summary for rainy seasons 1999/2001 at Tsholotsho and Gwanda South
Location |
Year |
Month |
Daily rad’n |
Max temp |
Min temp |
Rain total |
Tsholotsho |
1999/00 |
Dec-Mar |
20.7 |
28.6 |
15.5 |
340 |
2000/01 |
Dec-Mar |
19.8 |
26.6 |
15.3 |
590 | |
Gwanda South |
1999/00 |
Dec-Mar |
29.1 |
19.4 |
1405 | |
2000/01 |
Dec-Mar |
30.3 |
19.4 |
550 |
Enterprise and whole-farm budgeting was evaluated with @RISK. An enterprise budget was constructed for each FYM-aided production practice. Budgets were constructed using yield and input-output coefficients from survey data, on-farm experiments, and yields predicted by APSIM. Product values were obtained from government and farmer union sources, and input prices from suppliers.
Table 2. Analysis of quality of goat and cattle FYM from farms in Tsholotsho and Gwanda South in 1999-00 and 2000-01 – direct from kraal or composted by heaping/covering
Year |
Manure type |
Total N |
Total P |
Total C |
C/N |
NO3-N |
Olsen P |
1999-2000 |
Cattle from kraal 1-20 yr (mean of 4) |
10.8 |
1.4 |
168 |
15.5 |
1230 |
143 |
Cattle heaped/covered (mean of 4) |
11.0 |
1.2 |
167 |
16.2 |
1820 |
154 | |
Goat from kraal 1-3 years (mean of 3) |
21.9 |
2.2 |
281 |
12.9 |
1180 |
2430 | |
Goat heaped/covered (mean of 4) |
16.6 |
1.4 |
210 |
12.9 |
2670 |
1890 | |
2000-2001 |
Cattle from kraal, age unknown (mean of 2) |
10.1 |
3.1 |
134 |
13.0 |
509 |
182 |
Cattle heaped/covered (mean of 3) |
7.3 |
3.2 |
111 |
15.5 |
661 |
512 | |
Goat from kraal, age unknown (mean of 3) |
18.7 |
7.6 |
235 |
12.7 |
2460 |
747 | |
Goat heaped/covered (mean of 3) |
9.8 |
1.5 |
130 |
13.5 |
917 |
481 |
Total N and P concentrations vary, and generally higher in goat than cattle FYM. Organic C is also higher in goat than cattle FYM, and when <150, there may be greater inclusion of mineral soil. C/N ratio ranges from 13-17, and variation may be influenced by charcoal or lignin-derived material. High nitrate values indicate N mineralization while in the kraal, or during composting, and that any leaching of nitrate is incomplete. Accumulation of mineral P also occurs in the kraal or during composting.
We report from 4 farms in 1999-00 and 3 farms in 2000-01 where farmers tested goat and cattle FYM, and the effects of 3 pre-treatments, namely preparation by (1) storing the FYM in uncovered heaps, (2) storing in heaps covered with a layer of soil, and (3) storing in a pit covered with a layer of soil. FYM was taken to the field and applied in December prior to seeding. The experimental design was one or two FYM types (cattle and goat), two rates of FYM (0 and 5 t ha-1), 3 treatments of the FYM, and 2 replicates. The experiment and the surrounding area were sown by the farmer using sorghum (cv Macia) in 0.9 m rows and later thinned to 0.25 m between plants within rows. At maturity, the researchers and farmers harvested the experimental plots, and researchers determined grain yield for all treatments. In some cases, the residual value was estimated in the following season. Farmers separately evaluated the plots by observation prior to harvest. Farmers were positive about the effect of FYM type and pre-treatment, but harvested yields showed little difference between cattle and goat FYM (Table 3) or between FYM treatments, except for heaped-covered FYM in 1999-00 (Table 4).
Table 3. Evaluation of cattle and goat FYM for sorghum yield (t/ha) – mean from Gwanda South
Season |
No. of farms |
Control (no input) |
Cattle FYM |
Goat FYM |
1999/00 |
4 |
1.00 |
0.94 |
1.22 |
2000/01 |
3 |
1.05 |
0.99 |
0.98 |
Mean |
1.03 |
0.97 |
1.10 |
Table 4. Evaluation of FYM methods of preparation for sorghum yield (t/ha) – means from cattle and goat FYM in Gwanda South
Season |
No. of farms |
No input control |
Heaped covered |
Heaped uncovered |
Pit |
1999/00 |
4 |
1.00 |
1.34 |
1.00 |
0.91 |
2000/01 |
3 |
1.05 |
0.97 |
0.92 |
1.06 |
Mean |
1.03 |
1.15 |
0.96 |
0.99 |
On Johnson Nkomo’s farm in Gwanda South, on a Chromi-Leptic Cambisol (WRB) site, we compared goat and cattle FYM, and evaluated 3 pre-treatments. This carefully managed site showed good response to FYM (Table 5). Since we were constrained to two years of trials, the APSIM model (McCown et al. 1996) was used to simulate sorghum yield for 1990-01. Outputs were compared with sorghum performance with different treatments in the same seasons, and were in the range of the field values. Encouraged by comments from farmers, we used simulation to assess climatic risk of soil fertility inputs (Table 6).
Table 5. Evaluation of cattle and goat FYM with 3 methods of preparation for sorghum grain yield (t ha-1) on Johnson Nkomo’s farm in Gwanda South
No input |
Cattle |
Goat | |||||
Season |
Heaped covered |
Heaped uncovered |
Pit |
Heaped covered |
Heaped uncovered |
Pit | |
1999/00 |
1.54 |
1.97 |
1.58 |
1.50 |
3.27 |
2.26 |
1.46 |
2000/01 |
1.83 |
2.07 |
2.82 |
2.08 |
1.53 |
1.81 |
1.85 |
Mean |
1.69 |
2.02 |
2.20 |
1.79 |
2.40 |
2.04 |
1.65 |
Table 6. APSIM simulation of FYM with methods of preparation for sorghum grain yield (t/ha) on Johnson Nkomo’s farm in Gwanda South – means of treatment groupings.
Season |
No inputs |
Uncovered |
Covered |
Pit |
Cattle |
Goat |
Mean for 11 yrs |
0.52 |
0.88 |
0.87 |
0.62 |
0.73 |
0.94 |
Range for 11 yrs |
0.00-2.89 |
0.00-2.93 |
0.00-2.94 |
0.00-2.92 |
0.05-2.92 |
0.00-2.95 |
Failures per 11 yrs |
7 |
2 |
1 |
1 |
0 |
1 |
We used 3 farms in Tsholotsho and Gwanda South - S Mlambo (pelli-eutric vertisol (WRB)), B Moyo (eutric-aridic regosol (WRB)), and T Moyo (chromo-leptic cambisol (WRB)), a range of clay to sandy soils. There were 3 FYM inputs (0, 5 and 10 t ha-1) with 2 replicates. FYM was applied in December prior to sowing. Sorghum was sown by the farmer in 0.9 m rows, later thinned to 0.25 m between plants. At maturity, researchers and farmers harvested the plots, and researchers determined yield of all treatments. Farmers evaluated the plots by observation. The simulation package APSIM was used to simulate sorghum growth and yield for 1990-2001 using weather station and soil data. Outputs were compared with sorghum performance of the treatments, and used to assess climatic risk of soil FYM inputs.
Table 7. APSIM simulation of effect of 3 levels of FYM inputs, and 3 levels of N fertilizer on sorghum grain yield (t/ha) on three farmers’ fields in Tsholotsho and Gwanda South.
Farmer |
Soil type |
Zero FYM |
FYM 5 t/ha |
FYM 10 t/ha | |||||||
0 N |
9 N |
18 N |
0 N |
9 N |
18 N |
0 N |
9 N |
18 N | |||
B Moyo (Tsholotsho) |
Sand |
Mean |
0.67 |
1.12 |
1.45 |
0.95 |
1.28 |
1.45 |
1.10 |
1.42 |
1.44 |
Failures |
6/11 |
0/11 |
0/11 |
0/11 |
0/11 |
0/11 |
0/11 |
0/11 |
0/11 | ||
T Moyo (Gwanda South) |
Loam |
Mean |
1.01 |
1.47 |
1.81 |
1.28 |
1.75 |
1.95 |
1.53 |
1.85 |
2.34 |
Failures |
5/11 |
0/11 |
0/11 |
1/11 |
0/11 |
0/11 |
1/11 |
0/11 |
0/11 | ||
S Mlambo (Tsholotsho) |
Clay |
Mean |
2.83 |
3.37 |
3.47 |
2.95 |
3.57 |
3.69 |
3.20 |
3.36 |
3.94 |
Failures |
1/11 |
0/11 |
0/11 |
1/11 |
0/11 |
0/11 |
0/11 |
0/11 |
0/11 | ||
All soils |
Mean |
1.50 |
1.99 |
2.24 |
1.73 |
2.20 |
2.36 |
1.94 |
2.21 |
2.57 | |
Total failures |
12/33 |
0/33 |
0/33 |
2/33 |
0/33 |
0/33 |
1/33 |
0/33 |
0/33 |
At the sandy site of B Moyo and the loamy soil site of T Moyo, there were strong responses to inputs of modest amounts of N fertilizer, but only modest yield increase with FYM input (Table 7). Both fertilizer and FYM reduced the frequency of crop failure. At the S Mlambo clay soil site, there was a small trend towards FYM response (Table 7), and a strong response to N fertilizer. Risk of crop failures in dry years was less in the clay soil, and both FYM and N fertilizer reduced the risk of failure.
This work showed that FYM gave suggestions of responses, but results were variable even though farmers’ evaluations were positive. APSIM modelling indicated that without inputs, yields were low and there was high risk of crop failure. With added FYM, there were fewer crop failures, and about 50% higher yields. Both goat and cattle FYM were effective. Local farmers confirmed the idea that FYM had improved sorghum yield and that there was no ‘burning’. Simulations were useful in discussions with farmers who were patient with our efforts at simulation. Also farmers know that they need improved record keeping.
Rusike et al. (2003) evaluated the returns and risks above fixed costs for sorghum for 11 seasons, 1990/1 to 2000/1, for alternative soil fertility inputs, and for 3 different household categories - male-headed households (resident husband, more labour and use of draft animals), de facto female-headed households (absent husband – intermediate resource, better access to cash), and de jure female-headed households (most resource-constrained). For FYM inputs, each category performed similarly, with good returns from sorghum with kraal or pit FYM as seen for de jure female-headed households (Table 8). Returns were higher at the wetter Tsholotsho site. Higher returns came with higher risks, and lower expected returns with lower risks. Surveys show that most farmers prefer maize even where conditions favour sorghum and millet. They grow some sorghum and millet to diversify risk and as insurance against complete crop loss. Two-thirds of households face regular food deficits because of crop loss. Further simulation has suggested that small inputs of fertilizer, composted FYM, and FYM-fertilizer have potential for de facto female-headed households, legume rotations in male-headed households with draft animals, labour and land, and small inputs of fertilizer and legume intercrops in the de jure female-headed households.
Table 8. Expected returns and risk (Zimbabwe $ ha-1) of sorghum FYM management for de jure female-headed households in Gwanda South and Tsholotsho, 1990/1-2000/1.
Household type |
Gwanda South |
Tsholotsho |
|||
Activity |
Return |
Risk |
Return |
Risk | |
De jure female-headed |
Sorghum + kraal FYM |
-13023 |
3668 |
-7821 |
4410 |
Sorghum |
1704 |
8129 |
5419 |
6018 | |
Sorghum + pit FYM |
122 |
6398 |
6405 |
8531 |
Farmer surveys provided information about farmers’ knowledge and adoption of technologies (Table 9). Most farmers who hosted trials were adopting some ideas from the sites, and about half of farmers not hosting trials were adopting some ideas from trial plots. Most popular methods included heaped/covered FYM, and pit-composted FYM. Constraints to adoption were erratic rainfall and drought, lack of animals, ‘burning’ of crops by FYM, insufficient FYM, and lack of knowledge.
Table 9. Knowledge/adoption, and why not use FYM, Tsholotsho and Gwanda Sth (% respondents)
1998/99 |
2002/03 | ||||
GwandaS |
Tsholotsho |
GwandaS |
Tsholotsho | ||
Changed farmers’ knowledge/practice |
Know FYM |
na |
na |
98 |
99 |
|
Use FYM in surv yr |
3 |
20 |
20 |
15 |
Reasons for not using FYM |
Burns the crop |
62 |
7 |
20 |
2 |
|
No perceived benefits |
22 |
40 |
15 |
24 |
|
Not enough FYM av |
9 |
26 |
5 |
18 |
|
Lack of transport |
3 |
19 |
1 |
7 |
|
Lack of labour |
0 |
0 |
3 |
13 |
|
Lack of knowledge |
0 |
0 |
6 |
1 |
|
Low rainfall |
0 |
0 |
31 |
13 |
Declining soil fertility, and low and erratic rainfall are major constraints to increasing smallholder productivity in communal areas of semi-arid Zimbabwe. Improved soil water and nutrient management is needed, but in the present economic climate, it is not feasible to purchase fertiliser for sorghum. Small quantities of FYM have potential for higher rainfall areas. On-farm research, simulation, and surveys have led to adoption of FYM in all household categories. To help investment, we suggest extension services escape from ideal recommendations, and offer a basket of options. Extension services need to link to marketing to help adoption of fertility technologies for cash, although fertilizer is not yet an option on farms on communal lands for non-cash crops. Legumes will improve soil, are marketable, and provide incentives for farmers to adopt technologies.
McCown RL, Hammer GL, Hargreaves JNG, Holzworth DP and Freebairn DM (1996). APSIM: A novel software system for model development, model testing, and simulation in agricultural systems research. Agricultural Systems 50, 255-71.
Rusike J, Dimes J and Twomlow S (2003). Risk-return tradeoffs of smallholder investments in improved soil fertility management technologies in the semi-arid areas of Zimbabwe. IAAE Mini-symposium on Soil Fertility and Food Security for the Poor in Southern Africa: Technical, Policy and Institutional Challenge, Durban, South Africa, August 16-22, 2003.