Conservation agriculture: latest experiences from Zimbabwe

In this blog, I look at what happened, based on reflections from our field sites across the country – from Chikombedzi in Mwenezi in the far south, to Matobo in Matabeleland South to Masvingo and Gutu districts to Mvurwi in the north.

Across our sites, even in the resettlement areas where there are larger land areas, the uptake has been impressive. According to our informants (mostly agricultural extension officers living in the area), it is lowest in the tobacco farming area of Mvurwi (around 50%) and high in the poor sandy soil areas of Gutu/Chatsworth (over 90%) and Wondedzo (about 80%), as well as in the drier areas of Mwenezi and Matobo (80-90%).

But what Pfumvudza actually is in the different sites varies. Across the sites different packages – mostly maize, but also soya and sorghum – were offered. But there was a huge range of different seed varieties delivered. Some proved excellent, others less so. And the timing of the deliveries varied too. Some were available before the early rains, allowing dry planting, others arrived late, missing the plentiful early rains and hitting the mid-season drought that affected many sites. In addition to variations in types of seeds there were different levels of provision of fertiliser (compound D and top dressing), with many farmers complaining that this was inadequate. All these factors had a big effect on the outcomes of the programme.

Comparing Pfumvudza and conventional approach: a very rough assessment

In the last few weeks as crops have matured, the team has done a visual assessment of the likely harvests in the Pfumvudza plots and in other fields. This is very rough-and-ready, and should not be taken as a definitive assessment, but it’s based on long experience of working in the areas, and in most cases with experience as trained extension workers. Within these averages there is of course wide variation, much to do with timing. Those who planted early and benefited from early rains did well, both on their Pfumvudza plots and in their other fields. The results (with all the caveats) are in the table below.

Except in Mvurwi and Wondedzo, the Pfumvudza plots seem to have yielded more than the conventional agriculture in the open fields, but of course only on very small areas. The yield levels in the main fields this year were actually quite good, including in the usually very dry areas of Matobo and Mwenezi, where average yields are usually below a tonne per hectare. Any assessment must take account of what is happening in home and out fields, hence the comparison above. In good seasons, the use of more extensive outfields is feasible, and many ploughed furiously in December when the rains arrived in earnest. Even though planting late, they did reasonably well.

Of course the inputs supplied may not have all ended up in the Pfumvudza plots; as in past free distributions free inputs are applied carefully across the farm, making any evaluation tricky. In terms of the overall volume of output, outfield crops under conventional systems across several hectares will far exceed those produced in the small 0.06 ha Pfumvudza plots. Even with higher yields per hectare, the plots provide only a small fraction to the total. This of course might have been different in a dry year, when outfield crops may fail completely, and small garden-like plots provide an important production safety net, and so any evaluation must look across years, with the above figures taken in context.

Farmers’ reflections reveal a complex story

So what happened on the ground across our sites? A number of themes emerged in discussion with farmers and the field research team, relating to the effects of soils, rainfall pattern, seed supply, labour and politics:

Soils. Different soil types make a big difference. In sandy soils, there have been complaints of leaching due to heavy rains. This was particularly the case in Wondedzo in Masvingo where sandy soils suffered through the incessant rains this season. By contrast, in some areas where there are heavier soils and farmers complained about pooling of water in the pits and waterlogging. This meant adaptation of the system, including the building of cross furrows and other drainage systems, noticed in particular in Mvurwi, where the Pfumvudza plots fared worse than the conventional farming areas.

Rainfall. It was an unusually wet season this past year, with good early rains, a gap and then later rains. The season was in three periods, and those who planted early and got inputs in time did well. However those who planted later had poor results. This was a pattern across all our sites. However the high rainfall particularly affected our northern site in Mvurwi, which would normally expect reasonable rains for crop growth. Here waterlogging and even algal growth along with a massive weed burden proved a big problem in the conservation agriculture plots. By contrast, normal drainage and the use of herbicides in other fields proved helpful, resulting in higher yields there. By contrast in the dry south, where drought conditions are more common high rainfall on rich, heavy soils proved a bonanza and both Pfumvudza and conventional plots did spectacularly (at least for these areas). Of course any agronomic system must be able to adapt to different conditions, as there is no such thing as a ‘normal’ year. Mixing different approaches within a farm may be an important way forward, rather than seeing Pfumvudza as ‘the’ solution.

Seeds. It was comments about seed varieties that dominated the discussions with farmers across the sites. The government programme had a challenge in gaining access to seed and too often it resulted in inappropriate seeds being offered to Pfumvudza farmers. For example in Wondedzo and Gutu Chatsworth, farmers complained bitterly about the poorly-performing Syngenta variety supplied for their Pfumvudza plots. The SC513 that they bought locally did much better. In Mvurwi, farmers refused to collect the seed offered under the programme, and much remains rotting in the stores. Instead, they used their own seed, which proved more effective. In Matobo, farmers were happy with the Pioneer varieties that were supplied and this was the same in Mwenezi where Seed Coop varieties were offered. Early planting on high fertility soils in these sites resulted in bumper yields on the Pfumvudza plots, and also good yields elsewhere.

Labour. The digging of pits was a requirement for receiving inputs. So last year resulted in a growth in demand for labour, especially to help older and infirm people. Young men in particular were able to get piece-work employment during the lockdowns of 2020 to dig pits. And some richer farmers also employed labour as it is very hard work digging a full plot for Pfumvudza farming. Those in Mvurwi resettlement areas, where tobacco farming dominates, argued that Pfumvudza is not for commercial agriculture, where you need larger areas and digging pits is impossible (although with the loss of many cattle due to January disease last year, many had to resort to this technique due to a lack of draft power). The local nick-name for conservation agriculture is ‘dig and die’ (diga ufe). Many still refer to the programme in this way, but the term is now used quietly, as today criticising the programme is definitely not encouraged given its political cachet.

Politics. Many farmers complained about the politicisation of the programme. In the past you had to perform what the NGO or development project wanted to get the inputs for conservation agriculture, but now it’s more elaborate given how Pfumvudza has become a party-led, state backed campaign. Many commented that this politicisation means that (as with command agriculture) that patronage politics are played out around the programme, and those not supporting the programme are deemed to be in opposition to the government and are victimised. Others expressed suspicions that the Pfumvudza programme was part of a larger aim of down-sizing farms in the resettlement areas. If it can be proven that good yields are achievable on a small plot, then subdivisions become more possible, they observed. Farmers argued that the programme should be solely under the control of the ministry, and not within the purview of politicians, councillors and party cadres. However, the offer of free inputs is not shunned, although many argued that earlier programmes, such as the Presidential Support Scheme that was not tied to digging pits, were more effective. Many farmers said they would not continue with the practice if there were no free inputs.

Agronomy in context

In sum, while appreciating the programme, farmers complained a lot about the poor seeds, the late delivery and the uneven provision of inputs. They argued strongly that a blanket approach to the whole country controlled centrally – with everything from seeds, to fertility inputs to plant population to the size of the pits – does not make sense.

The programme instead needs to be much better attuned to local circumstances, including learning from how farmers have adapted the system, and not hiding this from extension workers and others for fear of admonishment. One extension worker recalled being tackled by a group of farmers earlier in the season: “You are from government”, they said, “what sort of people are you? You give us rubbish seeds. Who is responsible for this?”. As a front-line worker in a hierarchical, centralised system, he had no answer and had to agree (quietly). The failure to adjust, adapt and attune of course undermines any technological intervention. Learning from failures is always important.

Agronomy is always site specific, making big generalisations about interventions and techniques very problematic, as many reviews of conservation agriculture have pointed out (e.g. here and here). Context matters. There is never a magic bullet for farming. It all depends. This is why a more rounded perspective – beyond the idea of single magic bullet intervention – is needed. This is the theme of the blog next week, which is the final one in this Pfumvudza blog series.

This post was written by Ian Scoones and first appeared on Zimbabweland

Thanks to the team from across the country for their inputs and to Felix Murimbarimba for coordinating and compiling