Pakistan's outdated crop yield forecasting system needs a revamp, says Ibrar ul Hassan Akhtar.
Like most developing countries, Pakistan is staring at the spectre of food insecurity, with its food production out of sync with population growth. The food availability scenario is further complicated by changing weather patterns with recurring severe droughts and floods that affect crop production.
For government officials trying to deal with food scarcity, these endemic problems are worsened by inaccurate agricultural statistics which lead to subjective decisions in dealing with food security.
A predominantly agricultural country with 174 million  people, Pakistan still relies on an irrigation network developed during the British rule over the sub-continent with few additions like Tarbela and Mangla dams along with link canals. At independence in 1947, the irrigated area was around 10.75 million hectares, which has increased to over 18 million hectares . The current crop reporting system of collecting agricultural statistics from over 22 million hectares of cultivated area and on crop acreage and yields, is outdated. It is also mainly based on land revenue records, and limited to Punjab province.
The country urgently needs to revamp its crop yields forecasting and estimation system, if it is serious about tackling its food security problems.
Village surveys no longer suffice
After independence,  Pakistan's agriculture statistics were collected during opinion-based field surveys, which provided back-of-the envelope estimates that formed the basis for agriculture policy decisions for resolving food security issues. With the passage of time and advancements in science and technology, this system began to lose its reliability. Consistent rise in population  and incremental migration to urban areas were other factors that skewed the statistics. This outdated system was managed in the past by agriculture extension departments at the provincial level, and the ministry of food, agriculture and livestock at the federal level.
In late 1970s, the provincial crop reporting services (PCRSs) came into being as independent organisation, and by the early 1980s, data collection techniques for both people and crops changed. With technical assistance from foreign agencies, a system based partially on remote sensing satellites came into being. Some projects based on aerial photography and spot satellite data were also executed but failed to establish their strength at the national level
Under a village master sampling system, statisticians randomly select villages in provinces to estimate crop area and yields, and extrapolate the village data to the district level. There are three timelines: 'first estimates', 'second estimates' and 'final estimates'. In first estimate, PCRS reports initial cropped acreage and opinion-based production forecast calculated on actual crop harvests from the previous year in selected villages. The second provides the final cropped acreage and yield estimates by sample crop harvesting for the selected villages, which are combined for district level estimates. The final estimates are derived from the cropped acreage figures of the revenue department and yield from the PCRS departments. For example, in wheat crops, the first estimate is made on 1 February, the second on 1 April and the last on 1 August of each year .
Potential gaps need to be addressed
The gaps are evident. It is not possible to cover the entire agriculture area within cropping season timeframe; yield estimates come from limited crop cuts data and without inputs from remote, inaccessible areas; variations in yields due to farmers' preferred crop management techniques are not accounted for and the final cropped acreage still comes from the non-agricultural revenue department. Generally, the agricultural statistics lag by three to four months after crop harvesting, leading to some irrational decision making on import and export of agriculture commodities, which is influenced more by the political and private sectors than by reliable crop statistics
Pakistan needs to move towards a combined system of remote sensing, geographical information system (GIS), its improved national agro-meteorological network and the irrigation network
Pakistan’s 30 agro-meteorological stations are inadequate to cover the entire cropped area, and there is not enough information on spatial pattern of such variables such as soil temperature, air temperature and localised rainfall as compared to rainfall recorded at a single meteorological station
Remote sensing and GIS technologies will greatly help access remote areas and monitor crop conditions. This will serve as an early warning system for the decision makers by identifying drought conditions and possible low-crop areas due to different environmental stresses such as pests and disease outbreak, or improper fertiliser application
The most significant benefit will be timely generation of reliable crop statistics before crop harvesting begins, with confidence that the estimates are correct in 95 per cent of the instances, and do not lag behind actual crop harvesting
The new system should be based on modern agriculture information-based systems and in keeping with international standards of generating agricultural statistics. It should also factor in global forecasts for crop yields. Only then can the country hope to have agricultural and food security policies in place that meet current ground realties.