IV International Symposium Engineering Management and Competitiveness 2014 (EMC 2014)
June 20-21, 2014, Zrenjanin, Serbia
Forecasted and simulated effects
of long term force-fields through the example
of the grain sector of Kazakhstan
Zoltán Varga
SzentIstvánUniversity,Gödöllő, Hungary
Abstract
Through similarity analysis, the yields of the next 12 years can be derived with high correlation even with only 8 years of previous yieldscreating a model of 81 countries (including Kazakhstan).The additive and multiplicative effects of the different types of years describedthe yields of cereals in frame of a special production functionincluding meteorological data in the case of Kazakhstan. The connection between regional climate and regional yields can also be modelled in a balanced way as well.The sensitivity of the climate factors is acceptable. The annual precipitation and/or its frequency can be the base of simulations related to the pay off of the irrigation in regional level. If - during the previous 8 years climate - better agricultural technologywas used, the non-climatic effects on the rate of increase of yields can be estimated. In case of the multiplicative modelling,it could bedetected that environmental factors can have real negative effects on yield level.
Key words: simulation,Kazakhstan, long-term yield outlook/estimation, climate change
1Introduction
Kazakhstan is one of the major wheat producer in the world. The climate has a significant effect on the country’s production capabilities. As a world leader producer their yields are affecting the world prices of the cereals. If a major producer will be out of the row, it will affect the world prices. Unfortunately the climate change will significantly decrease the possible production capacities, and it will increase the prices on the world market.
Because the Kazakh agriculture is in shortage of water, a decision support should emerge to examine the possibilities how to stabilize the production. This article deals with the sensitivity of the Kazakh cereal production in addition to the climate change.Although just a very few data is available about the Kazakh agriculture and other related data sets. So this gave the task and the question is it possible to handle the situation (the decision support) to provide a useful of forecast, and sensitivity analysis.
2Status of the Kazakh cereal sector
The primary producer regions of cereals are located around of the northern/north-central territories. Kostanay, North Kazakhstan, Akmola, part of Pavlodar and the north of Karaganda, including the northern parts of West Kazakhstan and Aqtobe. Here the topography is mainly flat and the production on rich and fertile chernozem and kashtan (chestnut) soils (c.f. Map Nr. 1.) accounts for approximetely 70% of the country’s total wheat output.
1. Map Soil map of Northern Kazakhstan, source:
The cereals grown here are mainly rain-fed. Northern Kazakhstan produces hard wheat because of the dry climate. Despite the vast area - 205,000 km2, greater than Germany, Poland, Italy, France and Spain combined – given for high quality and exclusively to arable production, yield is considered low by global standards. Although there are several risk factors, it can be stated, that there is a great opportunity to raise the production in Kazakhstan. The following (Nr. 1.) table shows the yield of cereals of Germany, Poland, Italy, France, Spain, Hungary and Kazakhstan in kg per hectar from 2008 to 2012.
1. Table Yield of cereals (kg per hectar), source:
Country / 2008 / 2009 / 2010 / 2011 / 2012Germany / 7119 / 7199 / 6718 / 6461 / 6900
Poland / 3217 / 3478 / 3389 / 3391 / 3585
Italy / 5353 / 5087 / 5441 / 5682 / 5328
France / 7289 / 7455 / 6970 / 6831 / 7524
Spain / 3581 / 2939 / 3231 / 3708 / 2886
Hungary / 5800 / 4715 / 4719 / 5103 / 3662
Kazakhstan / 1009 / 1249 / 804 / 1,688 / 950
Based on the Table Nr. 1. the following question is relevant to ask: which genetic potential can be assumed for Kazakhstan comparing to experiences of the cereal production in other countries/regions? With other words: can be estimated the yield of cereals for decades? These questions are relevant in order to support decisions about setting up e.g. irrigation equipments.
Several methods exist to estimate future yields. The most common procedure is the scenario making/planning, for long term, but this leave the future direction unanswered. Basically it creates an optimistic, realistic, and a pessimistic direction with their calculated effect on the given question. Another way is to monitor a field experiment to extrapolate its results to the given counrty for short term (Mkoga, Z.J. et. al., 2010). The greatest advantage of this method is to making possible to analyse other important inputs (e.g. the effect of irrigation, fertilization, tilliage systems, etc.). The third technique uses COCO (Completeness and Consistency initiative) to fill the data gaps, ‘Bayesian’ approachoes, and Hodrick-Prescott filters (Britz W. et al., 2005).
3Climate
The climate of Kazakhstan is typically semi-arid, with cold winters and warm summers. Located well outside the Aral Sea Drainage Basin to the south, where pressure for irrigation waters exists and droughts are persistent, risk to drought is frequent – two years in every five on average -, particulary during the May-August growing season when poor rainfall and heat often persist. Harsh winters are also a factor. Large scale irrigation does not exist. Consequently, reduced harvested area and yield losses/crop failure is not uncommon, leading to frequent, and sharp, year-to-year fluctuations, representing a considerable source of acute regional food insecurity, and international accessibilty and market supply. The high frequency of adverse production conditions reflect few strategies to cope with such variable growing conditions. Where agriculture is least modernized and farmers do not have access to improved cultivars, and eddective inputs, the consequences of inadequate moisture can be serve with knock on effects for the local rural economy. Irrespective of such enviromental risks, Kazakhstan continues to comprise a significant part of the easterns reach of the Eurasian wheat belt, and increasingly strategic component of international wheat supply, in particular with respect to the EU. Climate change also brings uncertanities to the prospects of sustainable and uninterrupted growth of wheat in this region, where it is considered vulnerable. According to climate change scenarios based on global climate modelling, further temperature increases with no significant gain on atmospheric precipation may lead to a drier climate. In parallel, the current climate zone boundaries (Map Nr. 2.) may shift northward, and wheat yields may be reduced more than by 25%. Such future risks should not be underestimated. Therefore simulation modells being capable to integrate the most influence factors to the cereal production are necessary to be able to plan long term decisions.
2. Map Average temperature in July, source:
4The problem
Imagining the previously adumbrated scenario, it has a high probability that the main grain producing regions of Northern Kazakhstan may lose their layer of soil with relatively high fertility rate (chernozem and kashtan) because of for example the stronger wind and/or water erosions. It will be necessary to introduce new technologies (chemicals, machines, etc), at least not to lose more yield. The simulation-oriented question is, if they increase the effectivness of agronomy (e.g. fertilizers or improved irrigation systems), how these factors will affect the yield. Is it even possible to moderate the effect of the climate change? In details:(Q1) Is there a strong correlation between the previous years and the future yield averages? (Q2) Is it possible to find a strong correlation between the climate and the yields? (Q3) Is there a strong correlation between the parameters of the simulation model (climate data) and the yield? (Q4) What is the difference between the regional production functions (cf. sensitivity or risk volume)? (Q5)Is it possible to estimate the impacts ofnon-climatic effects(fertilizers, irrigation, etc) on the production concerning yields? (Q6) Is it possible to identify the climatic factors which have negative effects on? (Q7) How often occure a positve effect related to the precipitation factors?Methodological background
Simulation modells of this article are derived based on similarity analyses. The partial model layers are hybridized and try to build the most consistent holistic data-universe about short and long term status of the cereal production of Kazakhstan. Due to preliminary project contributions of University Gödöllő, Hungary: e.g. SPEL-IDARA-CAPRI international project-series to ensure consistence databases for agricultural sector modelling(Pitlik, 2003), and PhD-dissertation of Bunkóczi (2013)to forecast e.g. yield of cereals. Several models exist(team-intern) for the forecast of yields(Batár, 2009, Szilágyi et al, 2013). The understanding of the climatic processes of Kazakhstan, should be startedat examining on years alreadyhave beenknown. When this correlation is given, and if reliable climatic forecast data is available, then the possible results of the climate change might have known better.
5Results
4.1Databases
Due to the separation of the former Soviet Union, most of the data for the current Kazakhstan is collected after the independence (December 25, 1991), from 1992. All the data of yields is downloaded from the Databank of World Bank, the regional data of yields were found on the webstie of the Kazakh Statistical Agency, Metherological data were partly found in English on a Spanish website. Unfortunately, the dataassets had several lacks. To avoid the mistakes come from these gaps, theyhadbeen denied, of course not in the meaning of calculate with them, but separated the full databases from the missing ones. In the end, two very different cities had been choosen:Almaty, which is the former capital andlying in different climate zone asKzyl Orda, which is situated in the desert zone of Kazakhstan.
For the climate data, mostly seven different attributes were available:Annual average temperature (°C) [T], Annual average maximum temperature (°C) [TM], Annual average minimum temperature (°C) [Tm], Total annual precipitation of rain and / or snow (mm) [PP], Total days with rain during the year [RA], Total days with snow during the year [SN], Total days with fog during the year [FG].
4.2Methodology
Similarity analysis (COCO: component-based object comparison for objektivity) was used for this project. This methodology gives the answare from one input table, which is an object-attribute matrix (OAM). The OAM basicly consists the experiences of the past in a way, that they can be interpreted as a function of the others. For example in the performance-price situation, the inputed performances and the prices regarding them will be compared. If a given performance costs a given price, then a linearly stronger performance will have a higher a price. Most of the cases we examine more than two possibilites, and that is why a linear performance increase, not just in this example, is very rare.
The details of a performance, or a production, are diverse. The theoratical sum of combinations of these diverse details is infinite, but it is possible to emerge from the past the most general combinations. This will be the input for the OAM, and this will “teach” the modell the patern it should follow and reproduce.
Consistency is an important expression for similarity analysis. It is the examination of the possible paterns the modell learn, and these system relationships should be non-contradictional. It is crucial to clear up the possible queastions the modell can answare. For example it is a risky question to ask will the price of a share, or the production, increase or decrease. The better approachshould be to examinethe given company’s performance andbalance, or in the case of production, the factors of it, like a puzzle. This is used as a control as well. The sum of factors should be equal to the total of categories.
4.3Models
For the closing,10 models had been created. 4 of the 8 additive models were constructed around Kzyl Orda, the other 4 were around Almaty, there was a multiplicative, and basically there was a model which consisted of 81 countries yield data.(Q1) This last model revealed that there is a strong (86% - where the number of objects is 81, it means a single pattern) correlation between the past and the future yields.(Q5) From the ‘Model of Countries’ it could be estimated the potential of the yields (further 1297 kg per hectare based on the given level of 1126) assuming the highest level of agronomy.
Important note: every item from the climate database has its own partial correlation. In some cases the sum of two of these items dominate the major infulence. (Q2-4KO) The Kzyl Orda model-group showed that the RA is themore significant for the yield increasement,not the PP. If the PP factor is not calculated with, then the sum of the TM and TM factors will bedeterminative, which is shown the best in one of thesemodels, when these climate factors had been compared to the regional yields, from 2003 to 2012, eliminating 2005 which was a data gap. In this case, the sum of TM and Tm factors has reached 50%, and the SN factor was the second in the list with 39%. It means that the regional yield is sensitive for the differencebetween the annual average maximum (TM) and minimum (Tm) temperature, and it has a positive effect on the grain yields in the region. But from 1992 to 1999, except 1998 (data gap), this ‘statement’ is not entirely true. The main factors become the same as before (the other two models also calculate with the 1992-1999 period) RA and T.(Q7KO) The frequency of the positve effect of irrigation is 71%. So it is worth to set up irrigation systems for the Kzyl Orda region.(Q2-4Ay)
The Almaty model-groupshowed lessdifferencethan the previous group. Instead of the two temperature factors, here the TM factor dominates all 4 models. So the hotter is in the region, the more yield they will have.(Q7Ay) The frequency of the positve effect of irrigation in the region of Almaty is 50%.The question should be emerged: will Kazakhstanreally lose with thehotter climate?(Q3) For these models, except the ‘Model of Countries’, the correlations between parameters and climate is 100% (7-9 years, 7 attributes). (Q6)In the multiplicative model, which way of thinking was the opposite of the previous ones, thenegative factors had been looking for not the positive ones. It came out that if there are less foggy days in a year, the yield of that year will be influenced negatively by this factor.
6Conclusion
The Kzyl Orda modell group represents the ‘climate changed’ regio, and the Almaty group is the one, where it is possible to compare the changes needed after a newer climate.The climate itself can not be changed, but the effects caused can be moderated. For example an intensive irrigation might be responsible for 71% of the non-enviromental impacts.Among the enviromental effects, the difference of the maxmimum (TM) and minimum temperature (Tm) factors will have the highest impact (50%) on cereal yields.As the climate change will be more intense, the difference between the maxmimum (TM) and minimum temperature (Tm) factors will be wider and wider. This would have a negative effect on the production, and without major intervention as well as on the world market.Comparing the Kazakh production to other countries, where the average yields are much higher, this possible effect might cause less damage than in Kazakhstan, where the yields hardly exceed 1000 kilogramms per hectar.The regional yields and meteorological data make already possible to handle with long term decision situation based on similarity analyses and their consistence-oriented multilayer logic.
References
Batár, Á.: Production functions in the precision agriculture, SZIU, Best Student Paper, 2009
Britz W. et al.: CAPRI Modelling System Documentation - Common Agricultural Policy Regional Impact Analysis, Bonn, 2005,
Bunkóczi, L.: Forecasts and efficiency calculations for agricultural modells, SZIU, PhD-dissertation, 2013 (draft version:
Catholic Kazakhstan: Average temperature in July, time of download: 2013,
Central & Eastern Europe, Caucasus, and Central Asia: Climatic Characteristics of the Provisional Zones of Priority Environmental Problems, time of download: 2013,
Dr Balaghi, Riad et al.: Agrometeorological Cereal Yield Forecasting in Morocco, National Institute for Agronomic Research (INRA), 2013,
Food and Agriculture Organization: Northern Kazakhstan, time of download: 2013,
Mkoga, Z.J. et. al.: Extrapolating effects of conservation tillage on dry spell mitigation, yield and productivity of water using simulation modelling, Physics and Chemistry of the Earth, 2010,
Pitlik, L. et. al.: New aspects of agricultural sector modelling, 2009, (
Pitlik, L. Online Tool for Similarity Analysis, 2012 (
Statistics Agency of the Republic of Kazakhstan: Productivity of the basic agricultural crops, time of download: 2013
Szilágyi L et al.: Yield forecast of agricultural products, SZIU, Best Student Paper, 2013
TuTiempo.net: Climate Kazakhstan, time of download: 2013
United States Department of Agriculture: Foreign Agricultural Service: Wheat Production in Kazakhstan, 2005, time of download: 2013,
World Bank: Cereal Yield (kg per hectare), time of download: 2013