Turkey - Agricultural development alternatives for growth with exports (Vol. 2 of 3) : Annexes 1-4

Report No. 4204-TU 4 Turkey Agricultural Development Alternatives for Growth with Exports (In Three Volumes) Volume II: Annexes 1-4 June 30, 1983 Europe, Middle East and North Africa Projects Department FOR OFFICIAL USE ONLY Document of the World Bank This document has a restricted distribution and may be used by recipients only in the performance of their offical duties Its contents may not otherwise be disclosed without Workd Bank authorization.  CURRENCY EQUIVALENTS US$1 = Turkish lira (TL) 2141/ TL 1 = US$0.005 TL 1,000,000 US$4,673 WEIGHTS AND MEASURES 1 kilogram (kg) = 2.20 pounds 1 metric ton = 1,000 kilograms 1 metric ton = 0.98 long ton I meter (m) = 1.09 yards I kilometer (km) = 0.62 mile I hectare (ha) = 10,000 m2 = 2.47 acres I decare = 0.1 ha = 0.25 acre I square kilometer (km2) = 0.386 square mile 1 liter (1) = 0.264 gallon ABBREVIATIONS AI - Artificial Insemination CBT - Central Bank of Turkey CGE - Computable General Equilibrium Model DSI - State Hydraulic Works agency within the Ministry of Energy EBK - Meat and Fish Organization, an SEE GDAA - General Directorate of Agricultural Affairs of MAF GDPPQ - General Directorate of Plant Protection and Quarantine of MAF GDSF - General Directorate of State Farms of MAF GDVA - General Directorate of Veterinary Affairs of MAF Hara - Animal Breeding State Farm, under the GDVA MAF - Ministry of Agriculture and Forestry ORKOY - Directorate of Forest Villages of MAF SEE - State Economic Enterprise SEKER - Turkish Sugar Factories, an SEE SIS - State Institute of Statistics SPO - State Planning Organization TASM - Turkey Agricultural Sector Model T&V - Training and Visit System of Agricultural Extension TCZB - Agricultural Bank of Turkey TEKEL - Tea & Tobacco Monopoly, a Parastatal TMO - Soil Products Office, an SEE TOPRAKSU - Land and Water Development agency of the Ministry of Village Affairs TSEK - Milk Industries Organization, an SEE TUBITAK - Turkish Scientific Research Council TYT - Wool and Mohair Corporation, an SEE TZ - The Provincial Extension Services of GDAA TZDK - Agricultural Supply Organization, an SEE YEM - Turkish Feed Industries, an SEE YSE - General Directorate of Rural Infrastructure of the Ministry of Village Affairs FISCAL YEAR Government of Turkey - January I to December 31 SEEs and TCZB - January 1 to December 31 1/ Exchange rate as of June 1, 1983.  FOR OFFICIAL USE ONLY TURKEY AGRICULTURAL DEVELOPMENT ALTERNATIVES FOR GROWTH WITH EXPORTS Volume II: Annexes 1-4 Table of Contents Page No. Annex I Patterns of Resource Use, Physical Performance, and Constraints (1968/70-1978/80) 1 Annex 2 Agricultural Exports and Marketing 84 Annex 3 Food Consumption and Nutrition 121 Annex 4 Agri-related Manufacturing Industries; An Overview 134 Note: The annexes are based upon information available to the study team as of June 1982. This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.   ANNEX 1 TURKEY AGRICULTURAL DEVELOPMENT ALTERNATIVES FOR GROWTH WITH EXPORTS Patterns of Resource Use, Physical Performance and Constraints I. Resource Base 1. Turkey, occupying a total area of about 78 million ha -- of which about 1.1 million ha are inland lakes -- is characterized by extreme geoclimatic diversity which permits a wide range of crops to be grown under both rainfed and irrigated conditions. Topographically, it consists mostly of an eastwards-rising central plateau (Anatolia) of 800-2,000 meters of elevation, bordered by mountains in the north and in the south, with fertile plains along the coast and inland valleys. Much of the country is mountainous or hilly, with about 17% of the area above 30% slope, a further 21% between 20-30% and 18% between 12-20% slope. Alluvial soils have the greatest importance for crop production; however, large areas of central Turkey are dominated by fine-textured medium to shallow calcareous soils of the reddish-brown and brown soil groups. 2. Turkey's climate is mainly determined by its physical features and varies widely across the country as well as within geoclimatic zones. The northern coastal zone bordering the Black Sea receives the highest annual rainfall (1,260-2,500 mm) which is distributed throughout the year and thus reduces summer heat; the southern and western coastal regions have a mediterranean-type climate with hot, dry summers and mild winters when most of the annual rainfall occurs. Being shielded from the maritime influence by coastal mountains, the interior regions (which comprise most of the country) are subject to continental influences, with a lower rainfall occuring mostly in spring and winter and more extreme temperatures than the coastal areas; cold winters in particular greatly limit the growing season and the range of crops. The central part of the Anatolian plateau has the lowest rainfall (200-600 mm) and hot, dry summers, while the northeastern part because of its higher elevation has a somewhat higher and more evenly-distributed rainfall, cooler summers but more severe winters. Also, the mountains create many different microclimates which vary with the run of the valleys and other topographic features. 3. In most of Turkey rainfall is scarce during the growing season even in normal years; important fluctuations in precipitation also occur from year to year, such that rainfall is generally unreliable throughout the country. Therefore, irrigation has been a major factor in increasing and stabilizing agricultural production. From the average annual surface water run-off of 181 billion mJ, it is estimated by the State Hydraulic Works agency (DSI)  - 2 - that about 95 billion m3 I/ are physically available for irrigation and domestic/industrial water supply; in addition, groundwater reserves are estimated to have an annual yield capability of about 9.4 billion m3 2/ Under the present state of technology, it is crudely estimated by DSI and TOPRAKSU (the Land and Water Conservation agency) that about 8.5 million ha of land could be irrigated. Some 3 million ha are presently irrigated, of which about 1.1 million ha are served by large-scale public irrigation facilities built by DSI, 0.9 million ha by small-scale facilities built by TOPRAKSU, and the rest has been privately developed under usually more primitive systems. Even on publicly-irrigated land, water efficiency and cropping intensity are on the whole rather low. II. Land Use Sources of Data 4. The main sources of data about land use in Turkey are: (i) the annual agricultural statistics of the State Institute of Statistics (SIS), issued under the title "Agricultural Structure and Production" with an abridged version called "The Summary of Agricultural Statistics"; (ii) the General.Census of Agriculture conducted by SIS in 1970 and in 1980 on the basis of large sample surveys; and (iii) a study issued by TOPRAKSU in 1978, "Land Wealth of Turkey - Utilization, Classes, Problems", based on field surveys conducted during the 1966-71 period. 5. Since the agricultural census of 1950, the State Institute of Statistics has divided Turkey into 9 agricultural regions (see map) based mainly on topographic and climatic features, which are major determinants of the types of farming systems. The main geoclimatic and agricultural characteristics of these regions are summarized in Table 1. 6. Agricultural Structure and Production. The annual SIS agricultural statistics give provincial, regional and national figures on area, production and yield of various field crops and fruit trees, as well as data on livestock numbers and production and agricultural equipment. However, the usefulness of these data is limited both by the estimating techniques and stratification employed and by the manner of presentation. The annual statistics do not distinguish between irrigated and non-irrigated land, give only national figures for fallow and forest areas, and do not give any indication on the extent meadows and pastures areas. The vegetable area is not broken down into different kinds of vegetables, and for certain crops which can be used either for grain or for fodder production (such as maize or vetch) only the total area is given without further stratifications by activity. These agricultural statistics are presently estimated by the national and provincial extension agents of the Ministry of Agriculture and Forestry (MAF) on behalf of SIS, and 1/ Of which, only 9% is presently used. 2/ Of which, according to DSI, some 4.4 billion m3 are presently used.  -3- are not based upon a scientifically designed sampling frame with systematic surveys. Weaknesses in these statistics are recognized, and it is the intention of MAF to assign one statistics specialist in each province, whose task will be specifically to collect data on agricultural production; this is planned to start in 1983 together with the related training, and could lead to more realistic estimates for agricultural statistics, although the main instrumentalities will continue to be eye estimation techniques and interviews without independent verification. Presently, it is generally recognized among other weaknesses that the total wheat area may be overestimated by some 10-20% (and the production even ore so, because of a joint, but slighter, overestimation of yield)!I; also there are inconsistencies between the production statistics of certain pulses and the corresponding export statistics, which sometimes have exceeded the estimated,production (e.g. lentils in 1980-81). While the absolute figures reported in the annual SIS agricultural statistics are thus to be treated with extreme caution, the relative changes from year to year are probably right in direction and, to a lesser extent, in magnitude.2/ 7. Agricultural Censuses. By law the State Institute of Statistics is in charge of conducting an agricultural census every 10 years to provide additional information not covered by the annual statistics, and a basis for periodic revisions. Censuses were conducted in 1927, 1950, 1963, 1970 and 1980. The 1970 census was based both on area and farm unit samples, and included in particular some 200,000 household interviews. The published results include detailed data on land tenure and land use by farms size and by region. However, some problems arise from the fact that the total crop area reported by this census (about 15.8 million ha including fallow - see Table 2) is well below the usually accepted figure. The 1980 census is based on a sample survey of 3,000 villages with interviews in each of them of the village head and 6 randomly selected fa m households; in addition, all large-scale farm units have been surveyed.3' The total area reported by the 1980 census (Table 2) is more realistic than the one reported by the 1970 censu./; the 1/ In an attempt to improve its wheat statistics, the MAF recently initiated a project to forecast, monitor and verify production estimates using Landsat imagery and aerial photography. The project is being financed partially with proceeds from Loan 1762-TU (Grain Storage Project). 2/ There may however exist in the SIS statistics a general tendency to underestimate the magnitude of recent changes. This is probably the case for pulses in recent years, which seem to have experienced sharper increases in area and production than shown in the SIS statistics. On the other hand, the substantial overestimation of wheat area and production has been consistent since the early 1970s. 3/ This was the case also in the 1970 census, where large-scale holdings were defined as the ones operating 100 ha or more. In the 1980 census, the definition of large-scale holdings varies among regions according to the productivity of the land; in western Turkey for example, a large-scale holding is one with more than 50 ha irrigated or 100 ha dryland. 4/ The figures reported for the 1980 agricultural census derive from preliminary tallies from the village head questionnaires. Processing of the household questionnaires had not completed during the Mission's field visit (mid-1982).  - 4 - results of the 1980 census, both for areas and production, can thus be used to a certain extent for comparison with the annual statistics.!' On the other hand, this makes its results much less comparable with the 1970 census and the changes between the two censuses shown in Table 2 must, to a substantial but unknown extent, correspond to more reliable survey results rather than to actual increases in areas. Also, the preliminary results of the 1980 census do not give any information on land tenure systems, and only very sketchy indications on farm size distribution (see Section VI below). The village headman questionnaire results does provide, however, a breakdown by size of village land, the usefulness of which is questionable. 8. TOPRAKSU Survey. The purpose of the extensive survey conducted by TOPRAKSU during the 1966-71 period was to provide a basis for a rational and ecologically-safe land use planning. It involved ground surveys and soil sampling of all the provinces of Turkey. It provides provincial data about land-use in broad categories (dry agriculture with fallow, dry agriculture without fallow, wet agriculture, gardens and vineyards, special crops, meadows, pastures, forests, scrubs, other lands). The main emphasis, however, is on soil properties and conservation requirements, with a detailed soil classification by province into 8 capability groups. It is the only comprehensive study giving information about the extent of land misuse and the changes required for a better utilization. 9. General Land Use. About 44% of Turkey's agricultural land_2/ is occupied by permanent pastures and meadows.3/ The 1980 census and TOPRAKSU's study give quite different figures for the total area of pastures 1/ In particular, the 1980 census reports a much lower wheat area (and hence production) than the annual statistics, which is in line with the general feeling that wheat area and production are overestimated in the latter; similarly, the 1980 census estimate of the sugar beet area is lower than the corresponding estimate of the annual statistics which is probably also exaggerated (see Annex 5). On the other hand, the 1980 census probably underestimates in absolute terms the wheat area and production, as well as sugar beet production. Therefore, although it certainly represents an improvement over the 1970 census, its results should still be treated with caution. As shown in Table 2, the 1980 census includes in particular a huge area (17.6 million ha) under the category "non-arable land", which is quite unlikely to be idle in reality, if one considers the population pressure; this category certainly includes substantial under-reporting. 2/ Defined here as including crop land, meadows and pastures. Crop land is itself defined here as including the area sown in field crops, fallow, vineyards and orchards and vegetable gardens. 3/ The pastures ("Mer'a") are the range grazing lands, which may vary considerably in quality as well as in accessibility, while the meadows ("Cayir") are more specifically defined as grasslands, usually located in wet lowlands, and hence not grazed but used for hay production. The meadows, which are usually quite productive, represent only about 3% of the total; they are privately-owned, while the pastures (97% of the total) are most of the time communal lands.  -5- and meadows,-' but agree very closely on their relative share in total agricultural land, for the whole country as well as for each region (Table 3). Pastures and meadows are most important in the North East, South East and Central East regions, where they contribute well above half of the agricultural land; they are still quite important in the remainder of the Central Anatolian plateau (Central South and Central North regions) and in the Black Sea region, but become secondary in the western and southern coastal areas (Mediteranean, Aegean and Marmara regions), where they contribute less than 30% of the agricultural land. 10. The remaining 56% of agricultural land represent the crop land, which is occupied by field crops (roughly 60%), fallow (roughly 28%), and vineyards, orchards and vegetable gardens (roughly 12%).21 The dominant system of cultivation in rainfed areas consists of a rotation including fallow 1 year out of 2, especially for the interior regions; as shown in Table 4, fallow is of lesser importance in the coastal regions, which benefit from higher rainfall (in particular, the Black Sea and Marmara regions) and are usually better equipped with irrigation facilities (e.g. the Mediterranean and Aegean regions, see Table 5). Use of Cultivated Area. 11. SIS annual estimates. The regional distribution of cultivated area3/ is displayed in Table 6. The results of the SIS annual statistics as detailed in Table 7 indicate that cereals are most important amongst field crops, accounting for about 82% of the sown area in 1978 - 19804/ (from 67% 1/ 14.2 million ha according to the 1980 census, 21.7 million ha according to TOPRAKSU. As explained in the note to para. 7 there probably is some under-reporting in the census. On the other hand, some of the areas classified as pastures by TOPRAKSU probably correspond to very low productivity and/or remote and relatively inaccessible rangeland; part of this 21.7 million ha is thus not used or not usable (while other parts are overgrazed). 2/ For 1980, total crop land is estimated at 28.4 million ha in the annual statistics, and only 19.7 million in the agricultural census. Part of this difference is probably due to some under-reporting in the agricultural census (see note to para. 7). The reported use of the crop land in percentage terms is the following: 1980 annual statistics 1980 agricultural census Sown area (field crops) 57.7% 63.2% Fallow 28.9% 25.3% Vineyards & orchards 10.6% 9.3% Vegetable gardens 2.8% 2.1% 3/ Excludes fallow. 4/ All figures in paragraphs 11 and 12 are three-year running averages, derived from the SIS annual statistics for 1978, 1979 and 1980.  -6- in the Marmara and Aegean regions to almost 92% in Central South). The main cereal is wheat (56% of the sown area - see Table 8), followed by barley (17%) and maYze (3.5% only, but particularly important in the Black Sea region). Industrial crops have the second position after cereals, and occupy on average some 7.5% of the sown area (from less than 3% in Central North and Central South to more than 24% in Mediterranean region). Cotton is the most important industrial crop by area sown (4% of the sown area on average) and is found predominantly in the Mediterranean and to a lesser extent is the Aegean region; tobacco is essentially concentrated in the Aegean region and to a much lesser extent in the Black Sea and Central East regions; sugar beet is spread all over the country but mostly in the three Central Anatolian regions and in the North East region, as well as to a lesser extent the Marmara region. Oilseed crops, which on average account for only 3.4% of the total sown area, consist mostly of sunflower which is concentrated in the Marmara region where alone it accounts for 24% of the sown area. Pulses (mostly chickpeas and lentils) occupy on average about 4.2% of the sown area. Tuber crops (potato and dry onion mostly) occupy about 1.6% of the sown area (from 0.5% in South East to about 3% in North East and Black Sea regions). Fodder crops (alfalfa and sainfoin) are still very little developed throughout the country (about 1.1% of the sown area), despite the importance of the livestock sector. 12. The area in vineyards, orchards and vegetables is of particular importance (i) in the Black Sea region, where nut trees (mostly hazelnuts) are estimated to occupy on area equivalent to about 75% of the area sown annually in cereals (see Table 7); and (ii) in the Aegean, Mediterranean and Marmara regions. All these three regions, and in particular the Mediterranean region, devote a more significant part of their land to vegetable production than any other regions of the country. The Aegean region is also characterized by the importance of olive groves, which occupy an area equivalent to about half the area sown in cereals; this region also produces figs and some citrus. The Mediterranean region is the main growing area for citrus, and also produces some olives and figs. The Marmara region is less important in terms of fruit production than the preceding two, but produces a wide variety of fruit (including olive and hazelnut) with the exception of citrus. Grapes are also grown throughout the southern provinces of the country, while apples and other pome fruit, although grown throughout the country, are especially important in the Central South region. 13. Comparison with other data. The allocation of cultivated area as derived from the SIS annual statistics for 1980 is compared with similar data from the 1980 agricultural census in Tables 9 (field crops) and 10 (fruits and vegetables). For the whole country, the sown area reported in the census is only about 76% of the area reported in the annual statistics. The difference however varies considerably among regions, since the census reports a higher sown area in North East, and a similar one in Marmara and Aegean regions as compared to the annual statistics; on the other hand in Central South the sown area reported in the census is only 50% of the one reported in the annual statistics. As far as the utilization of the sown area is concerned, the census attributes on average a larger share to oilseeds, industrial crops and pulses and a smaller share to cereals, tuber crops and fodder crops than the  -7- annual statistics doll (the national areas in sunflower, tobacco, cotton and pulses as reported in the census are even higher in absolute terms than the corresponding figures from the annual statistics). This however, also varies among regions: for instance, cereals are given a larger share of the sown area in the census than in the agricultural statistics for the North East and Black Sea regions; pulses are given relatively less importance in the census in North East, Black Sea, Mediterranean, Marmara and Aegean regions; tubers are given more importance in Central South and fodder crops in South East. 14. The vegetable area reported in the census is only 69% of the corresponding area reported in the annual statistics; it however reports a much higher area in absolute terms in Central South (228%) and a similar one in Black Sea and Central East, while the reported areas in the main regions for vegetable production, i.e. Mediterranean, Aegean and Marmara, are respectively 54%, 50% and 48% of the annual statistics. The total fruit area reported in the census is about 54% of the corresponding area reported in the annual statistics-2 (from 40% in the Black Sea region to 67% in the Aegean region). The census also attributes a relatively larger share to apples and citrus, and a much smaller share to nuts and grapes than the annual statistics do. 15. The fact that the census attributes a larger share of the sown area to oilseeds and pulses, and a larger share of the fruit area to apples and citrus than the SIS annual statistics may be seen as a confirmation of the fact that the latter tend to underestimate the magnitude of recent changes in the cropping patterns, since these categories of crops experienced faster growth rates than the other ones, (discussed below in paras. 27 and 32). On the other hand, however, the census is characterized by some substantial underreporting which may have affected the various crops unevenly (e.g. underreporting may concern mainly areas under cereals). Also, it is recognized that the 1980 census tends to underreport relatively minor crops because of its sampling procedure; this may apply in particular to tubers and fodder crops. Trends in Land Use During the Last Decade3/ 16. Total Cultivated area. During the period 1968-80, the total cropped area in Turkey expanded by only 0.3% p.a. On the other hand, the total 1/ Among cereals, the census also attributes relatively less importance to wheat and more to barley and maize. 2/ The SIS annual statistics report only the number of trees for each species and each region, but not the corresponding areas which have been derived by applying the assumed densities of trees per ha indicated in Table 11. The SIS statistics give only a global estimate of the total fruit area of the whole country (3,019 thousand ha in 1980) which is below the area arrived at by applying such densities (3,404.9 thousand ha); if this global estimate is considered, the census results are 61% of the SIS estimate (instead of 54%). 3/ Except where mentioned otherwise, all results are derived from the SIS annual agricultural statistics.  cultivated area in Turkey (excludes fallow) increased by about 1.5 million ha at an annual rate of close to 0.8% a year (see Table 11). Of this, it can be estimated that about 0.5 million ha at most correspond to a reduction of the fallow, while at least 1 million ha has been gained from other land (grazing land or forests). The increase in cultivated area has been realized mostly in field crops (+0.8 million ha) and fruit trees (+0.5 million ha) and to a lesser extent vegetables (+0.2 million ha); these last two categories however experienced the fastest growth rates in area, respectively +3.4% a year for vegetable and +1.6% a year for fruit, while the area in field crops increased only by +0.5% a year. As a result, fruit and vegetable together now represent 19.7% of the cultivated area as opposed to 17.5% ten years ago (Table 7). 17. These trends are very similar throughout the regions. The total cultivated area has increased everywhere, except in Central South where it has remained more or less constant; the highest increase has been achieved in the South East region (+2.8% a year). The relative share of fruit and vegetable in the total cropped area has also increased in all regions except in Central South (where it has decreased slightly, mostly because of an absolute decrease in the vegetable area) and in South East where it has remained constant; in the latter region, however, there was an important increase in absolute terms of the fruit and vegetable areas. 18. Field Crops. Among field crops, the relative share of cereals has slightly decreased, from almost 85% of the area sown in the country ten years ago to a little more than 82% now (compare Table 12 and 8; the absolute amounts of area sown, by crop and region, are displayed in Tables 13 and 14); this applies to all individual regions as well. Nevertheless, in connection with the overall increase in cultivated area, the total area of cereals in the country shows a slight increase of about 0.2% a year (Table 15), which corresponds to a more important increase in South East (+2.4% a year) and to a certain extent in Central North (+0.28% a year). Elsewheres, the absolute area in cereals has remained more or less constant or even slightly decreased (Central South, Central East). Within cereals as a group, the wheat area has increased the most (about 0.9% a year), while the area of barley remained more or less constant and the area of other cereals (rye and oats in particular and to a lesser extent maYze) decreased. As a result, wheat has increased its relative share among cereals; this still remains valid is all regions taken individually with the exception of Central South, where barley has taken more importance. It should be noted however that for the last 2-3 years (1979-81), a reverse trend for barley and maYze is shown nationally, which have tended to increase in area at the expense of wheat. 19. As opposed to cereals, all other major categories of field crops (pulses, industrial crops, oilseeds, tubers, fodder crops) have increased their relative share of area sown between 1968 - 1980. In particular, oil seed crops experienced the fastest growth rate in area (+2.8% a year), mostly through an increase of the sunflower area of about 4.1% a year on lands previously sown to wheat, while oils of less importance (including sesame and soybean) have decreased except groundnut; although these trends apply very similarly throughout the regions, this obviously concerns mostly the Marmara region, which alone accounts for about 75% of the national area of sunflower and 66% of the total area of oilseeds, and to a lesser extent to the Aegean region (about 11% of the area in sunflower and 13% of the oilseeds).  - 9 - 20. The national area in pulses has increased by 2.4% a year, mostly chickpeas (+7.0%) and lentils (+6.2%), while the area of other important pulses has remained constant (beans) or decreased (vetch). Here again, the trends are very similar among regions: the total area in pulses has increased everywhere except in the Black Sea region where it has decreased; among pulses, the area allocated to chickpeas has increased in all regions, and lentils in most. The growth rate in pulses area has been particularly high where pulses were already important (Aegean, South East, Mediterranean and Central East, with respective growth of pulses area of 2.4%, 4.3%, 5.3% and 2.2% a year). 21. Although the share of sown area in industrial crops has also increased slightly (from about 7% of the sown area ten years ago to 7.5% now), the pattern has been more mixed. The total cotton area has remained constant on average,1 (though with large annual fluctuations) so that in fact industrial crops now account for a relatively smaller share of the total (increasing) sown area 2n the two regions (Mediterranean and Aegean) where most cotton is planted; the cotton area has also increased in the South East region, however, which in 1980 accounted for about 5.5% of the national area in cotton. In all of the other regions, industrial crops now represent an increased share of the field crops, mostly due to significant increases of the sugar beet area (+8.4% a year on average) throughout the country, including regions (South East, Mediterranean, Aegean) where it has not been particularly important. The total tobacco area has been decreasing by about -2.3% a year, which corresponds to a substantial decrease in the Aegean region (-5.1% a year), which accounts for about 55% of the national area, offset by a constant (Black Sea) or increasing (Central East) area in the two other important producing regions (about 15% of the national area each). The tuber crop area (potato and dry onion) has been increasing on average by +1.4% a year; it has however remained about constant in all the four coastal regions, while all other regions, and particularly so the North East region, increased their area in tubers. The area in fodder crops increased by a high as 5.2% a year for the whole country, but still contributes very little (about 1%) to the sown area. 22. Fruits and Nuts. Among fruit trees, the already important share of nuts (mostly hazelnuts and pistachio) has increased (compare Tables 16 and 17; the absolute areas in fruits and nuts, by region, are indicated in Tables 18 and 19), with an annual average growth in area of about +3.3% for the whole country (Table 20). Most of this growth was located in the Black Sea region which represents more than 70% of the national area in nut-trees (mainly hazelnuts). The share of citrus has also increased (growth rate of area of about +3.7% a year), in particular lemon and to a lesser extent mandarine. The growth in citrus area was experienced almost exclusively in the Mediterranean and Aegean regions, (which contained, respectively, 79% and 19% of the national citrus area by 1980). The area of certain pome and stone fruit such as apples, peach and apricot also increased faster than the overall 1/ It shows however more important fluctuations from one year to the other than most other crops. 2/ These two regions account for about 92% of the national area in cotton.  - 10 - fruit area. On the other hand, the olive area (mostly located in the Aegean and to a lesser extent Mediterranean regions, which account respectively for about 75% and 16% of the national olive area) grew at a much slower rate so that its share in the total has decreased. The grape area has decreased slightly in absolute terms, while the fig area remained constant. 23. Vegetables and Tea. Lastly, the share of vegetables and tea increased slightly as a result of fairly rapid growth (vegetables grew by 3.5% and tea by about 8.0% p.a. - data in Tables 16-20) though the absolute amount of land in tea remains insignificant. The growth in tea area was concentrated in the Black Sea region, which accounts for almost 90% of the tea area, although some growth also occurred in the Northeast which is the only other tea-producing region. Although in absolute terms the vegetables area increased in all regions except the Central South, the largest share continue to be located in the Mediterranean and Aegean regions. The most rapid expansion, however, occurred in the Southeast (7.6% p.a.), whose share in total vegetable area nearly tripled during the decade, and now (1980) is about 11 percent. III. Evolution of Crop Yields and Production-! 24. Although a steady if modest growth in cultivated area was realized during the past decade, a striking intensification of production occurred in the period under review. This was manifested principally through increased yields which, for most major crops and some fruits, generally accounted for far more of the realized growth in physical production than did areal expansion (Table 21). Some intensification was also due to the suppression of fallowed lands during the decade, which, in itself - while not adding to the cropped area - accounted for up to a third of the expansion of cultivated area (para. 16). From this performance and TOPRAKSU's land capability studies (paras. 39-43), it is evident that the arable land frontier was reached in the early 1970s, and that future production growth must be fostered principally through further intensification of the existing cropped area. The discussion which follows is based upon the evolution of physical production and yields2. of various crops in the agricultural regions of Turkey, as these are detailed in Tables 22-33. The detailed annual growth rates of production and yields during the 1968/70 - 1978/80 period have been derived from the SIS annual statistics, and are presented in Tables 34-37. 25. Trends in Crop Production. During the 1970s, the total production of cereals increased on average by 4.2% per year, most being due to improvement in yields, (+4.0% per year). Among cereals, wheat experienced the fastest production growth (+5.2% per year), followed by barley (+3.8%), and to a 1/ Except where mentioned otherwise, figures in this section are derived from SIS "Agricultural Structure and Production". 2/ For field crops, the yields considered here are per hectare sown, while SIS usually gives the yields per hectare harvested. For fruit trees, SIS reports the total number of trees and the number of bearing trees, as well as the production; yields per bearing tree have been derived from these data.  - 11 - lesser extent maize and rice, while the production of other minor cereals decreased in absolute terms. Yields improved for all cereals, but especially for wheat (+4.3% per year) as a result of an expansion of HYV's and to a lesser extent maize (+3.7% per year) and barley (+3.6% per year). 26. The most substantial gains in cereals yields were observed in the Marmara, Aegean, Mediterranean and Central North regions, which all experienced a growth rate in composite yield of more than 4% per year. Elsewhere, the improvement in cereals yield has been relatively less important; and also (except in the Black Sea region) of lesser statistical significance particularly in the semi-arid North East and South East regions, because of severe annual fluctuations due to a more profound variability of weather conditions and limited access to irrigation. Higher gains in cereals yields in the Marmara, Aegean and Mediterranean regions are mostly explained by the introduction in the 1970s of semi-dwarf soft wheat varieties, capable of responding to high level of fertilizer use; such varieties could not express their full yield potential in non-irrigated, low-rainfall areas and therefore spread mostly in the western and southern coastal areas. These were followed up by the introduction of modern semi-hard varieties for rainfed areas, which substantially out-performed traditional rainfed wheat whenever rainfall was adequate and correct cultivation methods applied. Although relatively more modest, the yields achieved in other regions also show improvement (particularly the Central Anatolian ones), resulting from uptake of semi-dwarf soft wheat and improved rainfed varieties, fertilizer use and tillage practices as well as, to a lesser extent, herbicide use. Thus, on the whole, the wheat improvement programs implemented by the Government in the 1970s successfully intensified the cultivation of cereals lands throughout the country. However, mainly because of differences in agro-climatic conditions, the gap has increased between the more developed regions and the less developed ones. Average cereals yields in the southern and western coastal areas are now around 220% of those in the less favorably endowed North East region, as opposed to 180% ten years ago. 27. Although the pulses production has increased on average by 3.0% per year, the average (composite) yield has not increased significantly and most of the increase in production is due to area increase. The national production of chick peas and lentils, which presently account for around 55% of all pulses, has been increasing by around 7% per year, with no significant increase in yield. The production of dry bean and broad bean (about 28% of all pulses) has been increasing by about 2% per year, mostly through yield increases. The production of all other pulses has been decreasing either by a decrease in area or, in the case of cow vetch, by a decrease in yield. The most important increases in pulses production have occurred in the Mediterranean region (6.8% per year for all pulses, more than 10% per year for lentils and chickpeas) and in the Aegean region (4.0% per year for all pulses, around 7% per year for lentils and chickpeas); increases in production have also occurred in Marmara, Central East and Central North, while elsewhere the production has been stagnant (Central South, South East, North East) or decreasing (Black Sea). The Aegean region is the only one where a significant increase in the composite yield of pulses has occurred (about 1.7% per year) but area increase (2.4% per year) has been a main factor in the total production increase of this region. Differences in pulses yields between  - 12 - regions are much lower than for cereals yields (the composite yield in the best region, i.e. Aegean, is around 140% of the composite yield in the poorest region, i.e. North East). 28. Among industrial crops, cotton production has remained more or less constant, with however a slight but insignificant increase of yield.l/ The production of tobacco has increased by 5% per year, because of a yield increase of 7.3% per year which has more than offset the decrease of area of about -2.3% per year. Sugarbeet yields have tended to decline in all regions, but area increase resulted in production growth of about 6.8% per year for the whole country.2/ 29. The production of oilseeds during the last decade has been increasing on average by 2.9% per year, because of yield increases (1.5% per year) as well as area increases (1.3% per year). The national cottonseed production, which presently accounts for about 51% of the total oilseed production, has remained more or less constant. However, sunflower production (located mainly in the Marmara region) presently accounts for about 40% of the total oilseed production, as opposed to only 28% ten years ago. Increases in sunflower seed production during the decade have been above 6% per year, mostly due to area increase (4% per year) and to a lesser extent yield increase (2% per year). Most of the area and production increase for sunflower took place during the early part of the 1970s, when sunflower displaced wheat in large areas of Thrace; area expansion subsequently levelled off after 1972-73. Yield increases, though more modest, have been steady through the period. 30. During the last decade, the production of tuber crops (mostly potato) increased steadily by 4.6% per year, principally through yield increases (3.2% per year) combined with some area increase (1.4% per year). The growth in area mostly occurred during the 1968-74 period, while significant and steady yield increases were realized after 1974. Area growth occurred mostly in the 1/ Since 1972, cotton yields have shown dramatic fluctuations from one year to another, probably reflecting the incidence of white fly attacks and other diseases. The SIS statistics show that each year of poor yield was followed by a lower area in cotton the following year, and each good year by higher area; as a result, areas sown to cotton shows even greater fluctuation than yields, as opposed to the pattern reported for other crops. 2/ The sugar beet production is estimated quite accurately in the SIS statistics since all the production is marketed and processed by the Turkish Sugar Factories Company. However, the estimates of area, and consequently yield, are less reliable. The reported area may be exaggerated, because both the company's extension agents (whose performance tend to be measured by the area planted in sugar beet) and the farmers (who can get credit through the Sugar Company) have incentive to do so. As a result, yields may be underestimated and the decline shown in the statistics artificial. Also, during recent years sugar beet production has expanded in the South East and North East regions, where yields are lower than the national average.  - 13 - Central and Eastern Anatolian regions, while yield increases were most important in the southern and western regions, less so in the Central Anatolian regions, and not significant in Eastern Anatolia. Little change in tuber areas or yields occurred in the Black Sea region. 31. The total production of fodder crops (alfalfa and sainfoin) increased by 5-6% per year, but still remains minimal. No yield increases were achieved in any of the 9 regions. 32. For all major categories of fruit (pome fruit, stone fruit, citrus, nuts) production increases of 4.1% (nuts) to almost 6% (citrus) have occurred, by a combination of both area and yield increases, as shown in Table 21. Fig and