Excellent wheat yields achieved in successful project-sponsored farmer trials during the 2007/08 season, combined with current high demand and record prices for wheat grain in local markets, have created exactly the right ingredients for achieving the project objective of expanding rabi cropping in southern Bangladesh. With the project providing targeted technical and resource support to motivated local change agents (Proshika, WRC and DAE staff) at this time of high interest in wheat by farmers, the prospect of significant impacts from continued project investment is enticingly high. Confidence in benchmarking, recording and quantifying such impact is greatly enhanced by the associated efforts in livelihood analysis supported by the project, All in all, this project is well on track to surpassing its initial expectations.
Farmers averaged nearly 4 t/ha in the large field-scale trials at Barisal and over 3.5 t/ha on Bhola island. In Noakhali, farmer yields averaged 2.3 t/ha across sites impacted by varying levels of salinity. Preliminary analysis indicates agronomic practices and released varieties, as recommended by WRC, are appropriate but that very late planting (late December) may reduce yields. In the more saline areas of Noakhali, varietal screening trials identified germplasm with consistent yield prospects for this region. While these results were influenced by the soft season finish of relatively cool temperatures and helpful rainfall during the 2007/08 season, this is the fourth consecutive season of acceptable wheat yields and so it is reasonable to reflect some confidence in achieving consistent crop yields in the regions of southern Bangladesh.
Detailed interpretation of the 2007/08 experimental results is yet to be completed and is dependent on soil chemical and physical data which are currently being analysed. Once available, these trial data will be used to test the APSIM model.
A remaining challenge to the project is to identify and spatially map land types with (i) current shallow tube well irrigation infrastructure, (ii) no salinity and surface storages of monsoon rainwater suitable for supplementary irrigation, (iii) near-surface water tables or high soil water holding capacity sufficient for dryland cropping or (iv) poor quality ground water but high quality surface water sufficient for irrigations. Some progress has been made in using satellite imagery to estimate surface water storage. Preliminary analyses, for example, indicate that for a 300ha region centred on Hazirhat, in the Noakhali region, there are 9.7ha of water bodies identifiable. This equates to 194ML (assuming 2m depth of water bodies) which is sufficient to potentially irrigate 100ha of wheat, assuming half this water is accessible for irrigation. Concentrated effort to progress such spatial mapping will occur during June 2008 when Perry Poulton visits Bangladesh with this specific purpose in mind.
Clearly impressive in the project is the collaboration evident between WRC and Proshika, such that it was not possible for Australian collaborators to discern which individuals were employed by each organisation. The project employed two scientific officers, one in each organisation, in each region (Barisal, Bhola, Noakhali) and they appeared to really work as a team. This was also reflected in the attitude and voiced views of the respective organisational leaders. The results were apparent in very well run on-farm trials and field days. While the Australian collaborators may be vocal in designing the field program, its implementation is largely up to the local team and they have exceeded expectations.
Given the excellent results to date, the project is well on course to ramp up extension and training activities for farmers in the upcoming 2008/09 season.

High world wheat prices and the need to import more than three quarters of domestic demand has renewed interest amongst Bangladeshi farmers in wheat as a Rabi-season (dry season) cropping option. Over the past 10-20 years wheat production had lost favour amongst farmers due to reductions in yield resulting from increased disease pressures and the relative attractiveness of alternative Rabi-season crops, particularly Boro rice and vegetables. Historically wheat has been grown in the north of the country, however this project focuses on the south, a region which had been considered unsuitable for wheat production, until FAO funded research (2003/05) achieved yields in excess of 2.5 t/ha with new, disease resistant varieties grown using appropriate agronomic practice. Current research builds on this work exploring options to improve management of resources including irrigation and nutrients and to adapt farm management options to deal with specific local constraints. There has also been a broadening of research focus to include other Rabi-season cropping options, particularly pulses such as mungbean, and an increased emphasis on the extension of research findings to the southern Bangladesh farming community. Wheat yields of 3 to 4.5 t/ha have been consistently achieved over the 4 years of research although higher temperatures during grain fill and a lack of rainfall during the 2008/09 season reduced yields by around 20% across the south. This is however, a very positive outcome considering that crops were sown in late December, 1 month later than traditionally considered optimal, due to the late harvest of T. Aman rice.
The increased availability of short season (60-100 days), well adapted wheat varieties bred by the Bangladesh Wheat Research Centre (WRC), and mungbean varieties which mature in 60 days, have contributed significantly to southern farming systems, providing farmers with increased choice in Rabi-season cropping. Wheat research under the project is now focussed on manipulation of variables including time of sowing, crop nutrition and irrigation to optimise yield and resource use. Trials undertaken during 2007/08 indicated that one irrigation at 20 days after sowing (DAS) was the most efficient use of limited water resources, increasing wheat yield by 0.7 t/ha compared to three irrigations (with 100 kg/ha of nitrogen (N) applied). Whilst research to explain this result is continuing it is thought that the response relates to increased mobilisation of applied N and associated improved development of adventitious roots and tillers. N use efficiency also improved with the use of one irrigation with a linear response to applied rates from 0 to 100 kg N/ha. This compared to three irrigations where response plateaued at 66 kg/ha and dryland which plateaued at 33 kg/ha. Recommendations for wheat production in southern Bangladesh have been changed in light of these findings and now differ from those in the north where at least 3 irrigations are required for successful wheat production. This begs the question-where is the crop sourcing the water necessary to support the levels of production being achieved in the south?
Analysis of records of ground water level at key locations across the south suggest an under-utilised resource that, because the water table is high, has the potential to contribute significantly to wheat production through capillary rise. This effect is minimal further north in Bangladesh as the water table is much deeper. Modelling of wheat production at Bhola suggests that the mean contribution of the shallow water table to transpired crop water use is 62mm (SD=18.8mm) under irrigation and 79mm (SD=18.9mm) under dryland. This equates to a difference in mean simulated grain yield of 1.1 t/ha (SD=0.62 t/ha) and 1.4 t/ha (SD=0.38 t/ha) respectively. The importance of growing crops which are better able to utilise scarce water resources has not been lost on the farmers of Noakhali and Comilla districts who were forced to reduce Boro rice plantings during the recent season, or had crops die prematurely due to water shortages. Rice has a demand for at least 15 irrigations per season compared to wheat's 1 making it look increasingly attractive in a water constrained environment.
On-farm demonstration sites, established by the project, which double as seed multiplication sites (SMT) for the new elite varieties are critical to the expansion of Rabi-season cropping. Collaborating farmers are supported by 16 regional research officers employed by the WRC and the NGO, Proshika. In 2007/08 45 sites were established in three regions with each farmer provided with seed, fertiliser and technical support to grow wheat and mungbeans. Farmers were trained to store their seed and undertook to sell at least some to neighbours for sowing in the following season. This program was expanded to 87 sites across six regions in 2008/09 and has been successful in showcasing options for Rabi-season cropping, developing expertise within the farming community and providing seed for future sowing. The success of this activity is exemplified in the village of Babuganj where Nasima Begum grew 800 m of wheat in 2007/08 (3.7 t/ha yield) and then sowed the land to mungbeans (1 t/ha yield). In 2008/09 she expanded production to 1000 m of wheat (3.5 t/ha yield) and 2400 m of mungbean and profited from the sale of wheat seed to 8 neighbours and mungbean to 5. Nasima is now recognised as the local farmer expert on the production of wheat and mungbeans and is sought out by her neighbours for advice.
Quantification of available fallow land for Rabi-season cropping continues to be an important objective of the project. Constraints on accessing regional data for real-time in-season land use assessment has lead to the trialling of satellite imagery as a potential data source. Analysis of the Noakhali region for the 2006/07, 2007/08 cropping seasons demonstrate the potential to classify land use at small scale paddock level (< 0.2ha). Imaging analysis is used to identify and then quantify land remaining fallow with estimates of 10 to 14% of the land in the study areas (293 ha & 1592 ha) remaining fallow during January to March 2007. Differences between the generated data and estimates by the Dept of Ag Extension (6% fallow in 2005) made it necessary to ground truth these regions in 2009 to enable improved differentiation between fallows and similar land uses not well defined using satellite imagery. It is expected that this work will result in more accurate assessment of land use and determination of actual land available for expanded Rabi-season cropping.