(Agenda item 5: papers 3, 4, 5)
African countries contribute roughly 23% to the world production of root and tuber crops (primary crops). The main root and tubers crops produced in Africa are:
Cassava (for which Africa represents 53% of the world production), followed by Asia (29%) and South America (17%);
Yams (96% of the world production);
Sweetpotatoes (7%), the main producer being Asia (91% of the world production);
Potatoes (4% of the world production), the main producers being Asia (37%), and the rest of the world (55%);
Other root and tuber crops (70%), followed by Asia (20% of world production).
In terms of contribution to calorie supply, the importance of roots, tubers and derived products (all production included and converted into primary product equivalent) is small, compared to the contribution of cereals. The contribution of root and tuber crops to the world supply of calories is only 5% compared to 48% for cereals and 46% for other foods. In Africa, root and tuber crops contribute 14% to the calorie supply as compared to 51% for cereals and 37% for other foods, while in South America roots and tubers contribute 5% and in Asia only 4% of the calorie supply.
The estimation of production of root and tuber crops raises a number of problems, especially the estimation of the cassava production. First, the estimation of the area under cultivation of cassava depends on the variety. Some varieties of cassava can be harvested after one year; in this case, the estimation of the production is relatively easy. However, in many countries, the harvest can extend over three years or more, and therefore the estimation of the production is difficult. Secondly, the concept of the cassava production is not the same in all countries. The statistics provided depend on the way the cassava is consumed. In some countries, the cassava is consumed as fresh boiled roots, while in other countries it is consumed in the form of meal, after long processing. In some statistical publications, there is no specification of the data presented, i.e. raw cassava or cassava meal and the comparison of data thus becomes difficult.
Importance of root and tuber crops in Africa
In Africa, roots and tubers are generally grown in countries located in the Sub-Saharan zones, notably in Nigeria, Ghana, Côte d’Ivoire, Benin, Cameroon, Central African Republic, Democratic Republic of the Congo (D.R.C.), Tanzania, Mozambique, Angola, Uganda, Malawi, Madagascar and Rwanda. In Africa, roughly 40% of all root and tuber crops are produced by Nigeria, followed by the DRC (10%), Ghana (8%), Tanzania (4%), Mozambique (3%), Uganda (5%) and Côte d’Ivoire (3%).
Cassava is produced mostly by Nigeria (35%), followed by D.R.C. (17%), Ghana (9%), Tanzania (6%), Mozambique (6%), Angola (5%) and Uganda (5%). Yam is produced mainly by Nigeria (71%), Ghana (9%), Côte d’Ivoire (8%), Benin (5%) and the rest of Africa (7%). Sweetpotato is produced by Nigeria (24%), followed by Uganda (23%), Rwanda (10%), Tanzania (8%) and Madagascar (5%) and rest of Africa (29%). Potatoes (Irish potatoes) are produced by Malawi (16%), followed by Rwanda (8%) and Uganda (4%). The other countries in Africa produce 72%. Other root and tuber crops generally produced in Africa are taro and cocoyams. The largest producer is Nigeria (35%), followed by Ghana (14%), Cameroon (5%) and Côte d’Ivoire (3%). The other African countries produce 43%.
In terms of contribution of roots, tubers and derived products to calorie requirements, it should be noted that root and tuber crops contribute more than 600 calories per caput per day in the following countries: Angola, DRC, Republic of Congo (Congo-Brazzaville), Central African Republic, Mozambique, Ghana, Côte d’Ivoire, Rwanda, Togo and Benin. It is interesting to note that, although Nigeria is the largest producer of root and tuber crops in Africa, the contribution of these crops to calorie requirements is not the highest among the concerned countries (about 570). This can be explained by the great variety of sources of calories that are available in Nigeria.
In conclusion, it is noted that cassava is the most important root crop in Africa, but it is also the crop for which a sound methodology for estimation of the production does not exist.
The paper outlines some procedures that are followed in quantifying root crops in the national food balance sheets and the main problems that countries, especially in the Southern African Development Community (SADC) region, are encountering.
Quantifying root crops in the National Food Balance Sheet (NFBS) basically involves two main procedures. These include, firstly, standardization of the various forms in which data is collected on the concerned root commodity into one form. This procedure allows aggregation of the various forms of the commodity in which data is collected into a single figure. For example, cassava stock data may be collected in the form of wet cassava, dried cassava, cassava flour, or in the form of processed products. These forms of the root crop cannot just be added on a tonnage basis as they contain different food values.
Secondly, it is necessary to convert the root crop into some predetermined base commodity by the use of food energy values (FEV). This conversion makes it possible to compare and aggregate across commodities in the National Food Balance Sheet. In the countries where this conversion is done in the SADC region (for example, Malawi), maize is the base commodity and all other food commodities, including root crops are converted into maize equivalent.
The two main steps in quantifying root crops in the National Food Balance Sheet are relatively easy to follow once the data on the root crops are available and the conversion rates are known. However, it is the unavailability of reliable data on root crops, which is the main problem in all SADC countries. Methods of data collection currently in use are not reliable enough for a proper interpretation of the data in the balance sheets. In addition, country specific conversion rates are currently not available.
The general recommendation in this paper, therefore, is that FAO should help countries develop reliable methods of data collection on root crops. FAO experts may also initiate in-country studies/ experiments to generate the various country specific conversion rates.
Cassava is an important crop of the Thai economy. Its major production is in the Northeast. However, the total production declined during the past two decades due to a decrease in acreage. The demand for Thai cassava products depends on the overseas markets. Their policies will have an effect on the production, trade and prices of cassava in Thailand. Thus, forecast of the production is considered to be useful for policy makers to foresee and prepare some measures to cope with the changing situation.
The model construction was attempted at two levels, i.e. national and regional levels. Ordinary Least Squares was used to estimate the coefficients in each equation. Cobb-Douglas type was applied in the planted area equation while time series model was used in the yield one. Applying the model for the ex-ante forecast, the total production is expected to be almost the same in 2002 when compared with the previous year.
Weaknesses in the model still remain in terms of some specification errors. Thus, to make ex-ante forecast more useful, some policy variables should be added to reflect more the real situation.
In National Food Balance Sheets, the various food commodities are usually converted into a single base commodity and this is especially done in case of food cross-substitution. The need for such conversions was questioned and the option to convert all foodstuffs immediately into calories, fats and proteins was considered, taking into account the need to differentiate conversion rates according to Regions. The same breakdown could be used for both supply balance as for food requirement balance. The extent to which substitution can be assumed to take place in practice was considered to be heavily dependent on local and Regional food habits.
Regarding econometric models for forecasting food production, their usefulness was appreciated, but the omission of some variables could render the model less valid. The usefulness of time trend models was appreciated for the short-term, but the longer-term time trend models may loose predictive power. It was suggested that variables might be added that reflect aspects of policy in order that the effects of such policy be reflected. In addition, it was felt that multi-equation models might better reflect the existence of eventual interrelationships.
(Agenda item 6: papers 6, 7, 8, 9)
FAO is the only source of comparative statistical data across countries. Special effort has been made into the collection of reliable statistics for major export and food crops, in particular for crops such as cereals, which are easier to monitor through the marketing chain. But it has not been easy to deal with root and tuber crops when it comes to producing reliable data. This complicates the long term monitoring of trends and impact for root and tuber crops in solving problems of food security and in alleviating poverty. This may also affect policy analysis as well as the planning and management of food supplies.
Root and tuber crops especially cassava, potato, sweetpotato, yam, taro and tannia contribute significantly to human food availability in developing countries in Africa, Asia and the Pacific, and Latin America and the Caribbean. They are also used for animal feed and various industrial applications. Their broad agro-ecological adaptability and adaptation to marginal environments, great flexibility in mixed farming systems, ability to produce reasonable yields where most crops cannot and their capacity to provide high quantity of carbohydrates make them the basis for food security at household level. In developing statistical methodology for root and tuber crops, one must take into account some of the major agronomic aspects related to these crops such as long growth period for cassava, continuous planting and harvesting, mixed cropping and mixture of varieties, storage in the ground and post-harvest life and associated crop losses and damages, and leaf harvesting for human consumption.
Sampling units for estimating production at farmers’ field should not only represent the administrative settings of a particular country but they should take into the account the existing agro-ecological zones. Wherever leaf consumption and/or marketing are important, data collection and yield estimates should also include the weight of leaves. Baseline information collected through surveys should provide a better assessment of root and tuber crops productivity by farmer and production estimates should not overlook certain occurring events such as an outbreak of disease or pest and mass propagation and distribution of improved and proven varieties.
This paper is divided in 12 sections. 1) Estimating production should be based on known administrative units (AU) within which data can be collected by resident agents of government or other agencies; 2) Population of farms, farm-families, inputs, and tools commonly in use; 3) Terrain and time analysis of soil and weather by year by AU by period is important for each season so as to determine if the production data reflect what is expected for each AU; 4) Known fresh tuberous root yield from real farms (survey) over a wide area and for different conditions will produce a data bank of real yields against which other data can be cross-referenced and checked before being used in the development of the overall estimate of crop production; 5) Extrapolation of farming systems data: from local units usually realized by farmers so that the units are standardized and be repeatedly estimated whenever they are referred to in the data brought to the particular estimation process; 6) Crop life estimates: time series from planting to harvesting are important in estimating how much of the final yield of a mature crop would be obtained if harvested before full-term of crop development and maturation; 7) Deductions of losses arising from floods / damages can be used to bring the estimates in line more finely with the truth of measurement (accuracy); 8) Record of varieties and their potential yields by AU's to enable a better estimation of cases of mixed varieties especially where the ratios of varieties in farms are not known; 9) Crop area and planting time data, along the season over the 365 days of the year, so that the date of maturity of each batch of planted field can be incorporated in the overall output estimation process thereby avoiding over-estimation; 10) Estimates are constructed bit by bit in a realistic manner; 11) Recall surveys of how the season was from farmers especially because they do not keep regular records that could be used in the estimation process; and 12) Permanent estimation office where the job of continued development of accurate data on crop production can be undertaken.
In developing countries root and tuber crops (RTC) are an essential component of food security in rural as well as in urban areas. In addition, they are a source of cash for a lot of farmers, often the poorest ones. Among the RTC, yams play a major role, especially in West and Central Africa. Perusal of yield and production data raises certain issues related to its biodiversity (several species involved) and cropping technique patterns (e.g. single/double harvest). On the basis of experience gathered in research work carried out on yam-based cropping system between 1995 and 2000 in Benin, this presentation attempts to identify and to prioritize factors to be taken into consideration to develop a dataset allowing to estimate unitary yields as precisely as possible. Practical methods for evaluating the area cultivated to yam are also proposed.
The various constraints related to the collection of reliable statistics on yams (Dioscorea spp.) are reviewed. At this level, the study noted the importance of environmental factors and conditions as explanatory reasons for variation in production.
Based on a large-scale yam production study conducted in Côte d’Ivoire, the different stages in the "construction" of a development model for yam yields are described. The following factors were identified:
which constitute the most important components of yam tuber production.
The results of the study are critically examined and analyzed with particular emphasis on the methodological difficulties involved.
Any methodology for estimating production of root and tubers should assure integration of major agronomic aspects, such as growth period, continuous planting and harvesting, mixed cropping and mixture of varieties, post harvest losses, etc. In particular, due attention should be given to aspects of leaf picking, i.e. estimates of production of leaves and the effect of leaf picking on the growth and production of roots and tubers.
The advantages or otherwise of crop-specific surveys versus general-purpose surveys were raised and the various opinions on cost and time needed for data collection were discussed. It was also considered that farmers had a relatively short recall of past production, particularly in case of extended or continuous harvesting. Record keeping with the help of pictograms was an option to be considered. Some countries were attempting to use area estimates and assume the yield for cassava to be more or less constant for some years in order to estimate production. However, this required substantial local knowledge.
As to the accuracy of production estimates and in view of the existing large farming diversity, it was felt that it would be difficult to adopt a single approach. Wide periods of farm operations or continuous planting and harvesting may be dealt with by determining the mean of the time for each and organize data collection accordingly.
For the yam production estimates, it was agreed that plant density (or rather mound density) was an important variable, changing according to cropping pattern, but less so within an agro-ecological zone. Density could be determined relatively easily after planting. It could therefore be used in stratification for sampling. However, plant density should also be determined at harvest time, possibly in the form of counting the missing plants or empty mounds. The same applies to sweetpotato and other root crops.
After some discussion, it was clarified that always two variables had to be taken into account, i.e. both number of mounds and total planted area. In order to obtain a representative sample, it was considered that inter-farm variation was large and hence a large number of farms would be required rather than large sampling plots. However, the latter required further methodological investigation, but available experience showed the desirability of recommending a minimum plot size.
(Agenda item 7: papers 10, 11, 12, 13, 14)
The cassava production survey is a routine activity of the Office of Agricultural Economics (OAE), Ministry of Agriculture and Cooperatives. The information collected is used as a basis for formulating agricultural policy and planning. The survey is carried out at the national scale using sample survey techniques. It usually employs a stratified two-stage sample in which the villages are the primary sampling unit (psu) and the households are the ultimate sampling unit (usu). The data collection is done by OAE staff in each agro-economic zone through interview of the selected farmers in the sample. The data processing is carried out at the central office.
The main problems encountered in the survey are lack of timeliness and under-estimation. A further problem is for instance the use of a fairly complicated questionnaire, which makes it difficult for farmers to respond accurately. OAE has planned to change the data processing procedures, which now will be carried out in the field offices using notebook PC’s in order to facilitate making the provincial estimates.
In the Democratic Republic of Congo (DRC), cassava is the main staple food amongst root and tuber crops. It offers various products (fresh roots, cosettes, paste, chikwangues), which are consumed almost in all provinces, and, therefore, it forms an important part of the food habits of 70% of the population in the country.
Cassava is often cultivated in association with other crops. The need to assure food security explains the choice of cultivating cassava as dominant crop in mixtures with other crops since it constitutes the main staple food used by the entire population, due also to the fact that both roots and leaves are consumed.
With respect to producing statistical data, the assessment of cassava production has been made for a long time and the estimation of that production has gone through various stages of methodological development that has evolved over time.
Even though DRC has a long tradition of assessing crop production, the reliability of data on cassava production has always raised concern. Whilst the production of fresh roots might be known, that of the leaves is not, despite the fact that cassava leaves are ranked at the forefront of vegetable consumption by the Congolese population. It is therefore useful for statistics on cassava production to take into account the production of leaves.
The improvement of the estimation of cassava production and leaves depends on:
(i) an assessment of the fresh root production and leaves by regular visits to the producers (weekly or monthly) in order to capture the existing fluctuations in the harvested quantities of those products throughout all provinces.
(ii) Taking into account the local units of measurement used in transporting products from the harvest area to the village. A constant calibration of these units is needed.
(iii) Taking into account different varieties. Since the yields are different for the different varieties, it is not appropriate to apply a single average yield to cassava in general.
c) Paper 12: Report on a field experience “Estimation of production of cassava in the Bandundu (1987-1988) and Bas Congo (1988-1989) regions of D.R.C.”
Over the period 1987-1989, the Catholic University of Leuven (Belgium), in collaboration with the Ministry of Agriculture, Rural Extension and Community Development in the D.R.C. conducted large, in-depth surveys on agricultural production and marketing in Bandundu and Bas Congo provinces, as part of a larger study on food marketing for Kinshasa markets. The production surveys covered 3 products in Bandundu (cassava, maize, and groundnuts) and 7 in Bas Congo (the previous three plus cassava leaves, beans, plantains and rice). They were conducted over a full year, with weekly recordings of production in physical units (basins) on specially designed survey forms, with a large sample (1305 households in Bandundu, 1040 in Bas Congo). All physical units were converted to weights and the results were aggregated to cover the whole population.
The results thus obtained were compared with official production data. For Bandundu, the measured production of cassava was 37.6% higher than official data; in Bas Congo, it was 72.2% higher. Production was also much higher for all other crops, except plantains. For various reasons, particularly civil unrest in the 1990's, the official statistics, and in particular the methodology used, were not changed or adapted to take account of our findings.
Another aspect concerned a rapid and reliable methodology for estimation of agricultural production, which was tried out for several months. The method was based on recordings once a month by the field enumerator, of the number of basins harvested by a farmer. The results obtained compared very favourably with the results obtained by weekly recording by the village enumerators. This method holds a lot of potential for improving cassava production statistics and is cheap and reliable. It is based on the farmer's recall of the number of units (basins) harvested over a month.
Finally, the variances for all variables recorded in the surveys were calculated. Under certain assumptions, and for a given K (confidence level) and D (maximum difference between measured value and real value), this allowed a precise calculation of the required sample size. At 95% confidence level and 10% maximum error, the sample size for bitter cassava production in Bandundu was 103 households, and 240 in Bas Congo. For cassava, maize and groundnuts, the three main staple crops, and the income from the sale of the main cash crop, cassava, the required sample size for Bandundu was between 200 and 300.
It is concluded that actual food production in DRC is thus likely to be higher than official statistics indicate. The food situation would then appear better than official statistics indicate, although clearly still far from acceptable. Household budget food consumption data over the 1975-2000 period nevertheless also show a very poor nutritional situation, one of the worst in the world. But maybe also the household budget survey data are underestimated. In any case, the situation is alarming and better production statistics are an imperative for agricultural and food policy, for aid programs and for an improvement in the food insecurity situation.
The paper provides an overview of the concepts and methods recommended by FAO for the establishment of statistics on root and tuber crops. It presents FAO classification of the seven root and tuber crops defined by the Organization, with their common and Latin names.
The paper also reviews the definitions and various concepts of area, yield and production used in agricultural statistics as well as the related operational issues in the presence of some agricultural practices in traditional farming systems in Africa: continuous planting and harvesting, mixed and associated cropping.
The paper discusses the issues raised in estimating area, yield and production in the case of root and tuber crops with particular emphasis on cassava.
Finally the paper proposes some recommendations and elements of discussion, including: the need to address each root and tuber crop separately as the problems for estimating area, yield and production are different from one crop to another; the need to use stratification in agricultural surveys regarding root crops, using as stratifying variables, the production patterns, the varieties of crops, the need to conduct further studies and field tests to validate experiences undertaken in some countries. This could be done through a regional research project (3 years with a technical assistance component and pilot countries components).
Before the tragic events of 1994, the Agricultural Statistics Division (DSA) of the Ministry of Agriculture (MINAGRI) maintained a comprehensive database of agricultural statistics. The DSA was responsible for providing information on agricultural policy based on annual surveys of rural households. These surveys were conducted under the auspices of the Enquète Nationale Agricole (ENA).
These surveys (which were interrupted in 1994) were resumed in 1999 by the Food Security Research Project (FSRP) and the Agricultural Statistics Division of the MINAGRI. The FSRP/DSA began conducting agricultural surveys in 1999 using a national sample of 1584 households. The FSRP/DSA collects land use and production data on a seasonal basis (twice a year). The FSRP/DSA has 11 enumerators (one per province) as compared to 78 enumerators that the ENA had before 1994. The current sample size is also 26% larger than the one ENA used. Since the FSRP did not have as much financial resources available to the ENA as before 1994, it had to find a less costly but also accurate method to conduct the surveys. The most time consuming and therefore expensive activity of data collection was the area/field measurement. After considering various area measurement methodologies, the FSRP/DSA selected the P2/A (Perimeter Squared over Area) Methodology. This methodology minimizes time and costs.
The P2/A methodology is based on the unique and relatively stable relationship between a given field’s perimeter squared (P2) and its area (A) for a known form of field. Perimeter squared (P2) and area tend to vary together and in the same direction. In fact, the two are so highly correlated that a field’s perimeter could be used as a proxy measure of its area. This methodology further utilizes enumerator pacing around the field to measure the perimeter once the average length of the enumerator’s stride has been calibrated and recorded. This innovative methodology has allowed the FSRP/DSA to conduct surveys on a national sample with a limited number of enumerators and at a reasonable cost while generating statistically sound estimates.
Crop production is estimated using farmer recall for the season (6 months). Just after harvest, the enumerators visit the households and ask the farmers for the quantity produced during the season.
The actual methodology used for cassava statistics in agricultural surveys, conducted in Thailand tends to underestimate the production. Regular recording of farm production through farm bookkeeping as well as crop cutting may be alternative approaches to improve the estimates.
In some countries (e.g. DRC), cassava leaves contribute significantly to food security of populations, in addition to roots. Therefore, there is a need to improve statistical data on leaves in order to improve the assessment of the food security situation of the populations. In situations of subsistence farming, estimates of production are the most relevant variable for food security rather than area or yield. There is also a need to periodically evaluate area planted and yield. In situations of continuous planting and harvesting, frequent visits to record farmers’ production provide better estimates than one yearly measurement.
The optimum number of visits and ways of collecting data need to be further investigated for more cost-effective systems. Stratification should be used in order to reduce sample size and cost of surveys. Use of local measurement units requires proper calibration in order to provide metric equivalence.
It appears that various common local names with different meanings are often used for root and tuber crops. Therefore, it was proposed that in reporting statistics, Latin names be added to common names of crop plants.
The use of alternative methods of estimating area was discussed following the presentation of Paper 14. It was proposed that comparative studies be undertaken on cost-effectiveness of the alternative approaches (including use of GPS) as compared to traditional methods.
The appropriate institutional set-up for producing agricultural statistics was also discussed (centralized versus decentralized systems). It was recognized that there is no absolute best solution. Each country needs to adopt the most appropriate solution. Whatever system is selected, overall coordination of the different components of the statistical system is essential.
(Agenda item 8: paper 15)
Root crops especially cassava and sweetpotato are important food crops and contribute significantly to the food availability to the people of Malawi. Recurrent adverse weather conditions have contributed to the increasing production trends of root crops as compared to maize. Crop estimates conducted by the ministry of agriculture and irrigation do provide an indication of the availability of food per season. However, these figures do not take into consideration crop losses and results in over/under estimation of food availability. Major causes of crop losses in roots and tubers are pests such as cassava mealy bug, sweetpotato weevil, termites and green mite. Disease such as the cassava mosaic disease, cassava brown streak and cassava bacterial blight also do increase crop field losses while rotting, larger grain borer and cylas weevils are major causes of post-harvest losses of stored products. Crude methods of estimating crop losses are available although they are not utilized. Studies should be conducted to determine quantitative crop loss (pre- and post-harvest) estimates for each administrative or agro-ecological unit to improve estimation of food availability.
The various components and causes of losses were discussed. It was found that little experience or documentation existed on crop losses but the methodology being tested in Malawi was found encouraging. It was concluded that there was need for further methodological research on crop losses and also the concept(s) needed to be better defined.
The importance of information on crop losses for use in food balance sheets was recognized.
