USSR: The Impact of Recent Climate Change on Grain Production
; OFFICJaCoSE
Cop,
FOR OmClaVTUSE ONLY
USSR: THE IMPACT OF RECENT CLIMATE CHANGE ON GRAIN PRODUCTION
'} SUMMARY
;uring the past ISignificant fluctuation, which strongly aided grain production, occurred In the dimate of the Soviet grain belt. The severe droughtS may have marked the end of the favorable dimate trend Bideturn to the banner condition! of the.
"he climate fluctuation that occurred04 appears to be en anomaly. leaurUng in an increasing flow of maritime air from the North Atlantic tint generally caused
increased preclpltatkw, *
wanner winters, and
cooler summers.
The net effect was to move the moister northern climate southwardegrees of latitude, pushing back the desert and nearly doubling production in the new lands area of Kazakhstan.
;2otal grain production in the USSR increasedillion metric tons annually on thet is estimated that about half of the increase In grain production3 has been caused by the more favorablehe impact of climate, however, has varied greatly:
Most of the improvement occurred.
The Impact was largest in the southern fringes of the grain belt, east of the Urals. There was little or no impact In the normally moist areas such as the Baltics and Belorussia.
I. ihU raUKartoau trta* am ton cakaUra cat* im aawta laaaaaaSja 1 Data tatarfcaw aa* eto nllacttt* Manarm ofatM Miami ate a* afloarajaantftalhi mncaraca hrteml ta toM itM* aflact of cBaam. rtont.to* tat afreet of cBaaaMad M
It Is grown In areas with marginal rainfall, the Impact on springgrains production was much greater than the Impact on winter grains.
projections of average annual grain production haveased on different climatic assumptions:
Projection I- the "best"ssumes that the extraordinarily favorable climatere valla. This suggests an average annual cropillion tons.
Projectionecenta based on the climate. Accordingly, an avenge cropillion tons ta projected.
Projectionthe "wont" of the recenta based oo the climatend yields an avenge projected cropillion tons.
Climate cannot yet be forecasted reliably, but available evidence Indicates that the last projection is the most realistic one, putting the Soviet grain goalillion tons out of reach. An outputillion tons fallsillion tons short of estimated annual requirements. If the Soviets choose to cover the deficit by imports, purchases of foreign grain would match those following the disastrous grain harvests2ore likely, the Soviets, as they have done in the put, will reduce their requirementsearajor downturn in grain output by lowering their short-run goals for meat output and livestock inventories.
5 drought does not appear to be an aberration but partrier trend, which can be expected to occur with varying degrees of Intensity for some time to come. Duringhe Northern Hemisphere was cooling. This period was marked by hemispheric changes such as the Sahelian drought, failures of the Indian montoon, Increasing polar ice, end Increased rainfall in the Soviet grain belt, Recently, the cooling trend has reverted, and the Northern Hemisphere now Ii warming. Rami hive returned to the Sahel and India, and rainfall hai decreased in the Soviet grain belt. Changes of this magnl rude involve the exchange of large amounts of energy affecting climate over many yean. .
n' important assumption underlying our estimated average annual productionillion tons concerns the rote of "nonweather" fact on In changing grain yields. We have explicitly assumed that the trend In the usage ofd other yield-enhancingeferred to as "technological improveiB continuen increased priority for grain production could result In an acceleration in application of existing technology on grain at the expense of other crops during the next five yean.tep-up in the rate of growth In usage of fertilizer, for example, could alter upward the projection.
DISCUSSION
Introduction
The world's climate and tho possible effecthange in climate on food production is receiving increased attention. In particular, the drought and subsequent famine in the Sahelian zone of North Africa during thendas focused world attention on the implications of climate change. According to evidence gathered by climatologists, the Northern Hemisphere has been cooling since the. This cooling may have been responsible for the widespread failure duringf the rain-producing monsoons in the grain-growing regions that lie south of the tropical deserts.
In therain crop shortfall2 and subsequent massive imports drew attention to the potentially precarious situation someroducing countries In the North Temperate Zone might face because of climate-fluctuations. Bounded to the north by cold temperatures and to the south by deserts, the grain-growing region of the USSRigh potential for disastrous weather shou'd the boundaries of these unfavorable climates shift.
has been done to evaluate the effect that this climate changeon food production in the temperate latitudes. Thisiscussesof climate and climateses detailed meteorological datachanges in the climate in the USSR grain belt,stimates thethe climate change on grain production'
Background
The Nature of Climate
irruto It weatheronger lima Kiln. For example, diily mem tempmrure li used to dmeribo wither while mean temperature fordecade oried to chajicteru* dimate. Bothimages and both cha.tge with time. In region, where the water luppty Ii critically low, aoemlngly small change. In weather can have luge eontequencei on food production, and the differenceear droughteather phenomenon)limaticnly MmintJc
DUagreemrmt over the definition of climate, has frequently been over the length ofanging fromo more thanoceiavy to eiuWbh norma. For example, the World Meteorological Organluoon met the pan three complete dr.-adei to determine an area', climate, while the Department of Defense uteiear. In iti new worldwide meteorology data beac. None of thete definition. Ignore the Ice age. of the pait but rather aaiume tha" thehat the climate during the nextnean wfQ be much like thai of the iaitoeaii
In thli publication climate mean, weather averagede* or more. Thh definition make, no anumpllon about the liability of climate.
Is the weatheregion averaged over some period ofwhirling boundary of air that encircles the globe between the cold polar air
and warm subtropical air marks the location of the storm systems and weather fronts (seehe ske of the drcumpolar vortex, the dome of cold air covering the polar regions. Is related to the state of theemperature
:
WARM TROPICAL AIR
Tht Clrcumpolar Vortex, .
AVi -
and precipitation. As the Northern Hemisphere cools in the winter the circumpolar vortex expands, cooling the temperate latitudes and moving the hemispheric weather patterns southward. In the summer It con tracts, allowing wanner air to move north fromubtropics.
thai form along this boundary cause large masses of cold airsouthward into the warm subtropical oir, and warm air is forcedspinning storm systems that move eastward through the temperateanything from cloudy weather and drizzle to severe storms withdamaging winds. The shape, size, and number of these waves depend onof the polar vortex, tho temperature differences between the pole andand the topography over which the air flows.
in the USSR grain region is further controlled by thesystems that form over thj large land mass of Siberia and theOcean. Winters in the grain region are dominatediberiansystemorth Atlantic low pressure system causing the windfrom the southwest. This air originalcs in north Africa and southeasternis therefore dry. The storms bringing moisture from the North Atlanticthis gentle push to the north which makes it difficult forget into the southeastern portion of the Soviet grain region. Summersthe opposite pressure pattern,iberian loworth Atlanticair from the northwest* Because almost all of the water in thecomes from the North Atlantic, summer is the season ofegions. This air dries out as it moves cast and south, dropping lessprecipitation es it goes.
Climate aasslflcarton
Coram hati uvoufho-Ji ihe yen vrhOf bad Duficci chanrc tempera lure nptdly endire men differencesnmtr and winter. From January lo July Iho North Aiianitc >attei onlyelriui (CI while north-central Siberia variei moreC
The Koppen climatic clautflcaDon ayitem Ii dlscnaMd In CT.nto CKmtte, New
area's climate is commonly classified by average annualtemperature. It Is therefore possible to calculate valuesivenlimateesert, steppe, moist-continental, and the like -single year for any area. For example, Koppen, who derived the mostclassification system,oist-continental climate as onetemperature Is below freezing for at least oneC Tor at least one month,nnualannual evaporationhe entire Soviet grain belt qualifiesto temperature, but In many places evaporation potentialmaking the normalteppe rather than moist-continental
(see* If evaporation potential exceeds precipitationarge enoughteppeesert. Generally, excluding irrigation, small grains grow welloist-continental climate, with difficultyteppe, and not at allesert.
nan ii
Qimale Change
6, In tveh teajoni ntcami cannot originate: riven nowtng tnto thete rceloni loat. rather than tain, antcr.
Climate Is constantly changing and temperature Is usually used to measure the change. Although the difference between an ice ageild period may beew degrees Celsius In the average global temperature, tho result Is large climate changes over vast areas, Cooling Is associated with expansions of the polar vortexendency to southward migration of climate patterns. This shortens growing seasons in northern regions, brings rain to the northern edges of deserts, and dries the land south of the deserts.
Simple global climate fluctuations have complex local effects. As the polar vortex expands, lis wave pattern changes; air temperature and wind patterns affect ocean temperatures and currents that in turn affect the climate. In addition,
mountains alter the flow of winds; not only do the winds move norlh or south but also the precipitutlon patterns of these winds may change depending on the topography. Although the north torlimate changes may be Important, the east lorhanges are nearly always more important.
yearly, dally, and hourly variations In temperature andmuch greater thanear climaticong-termclimate may resultlightly cooler, more moist climatepecificit may also increase the variability in temperature and precipitation. As ahot, dry years may occur much more often than they didthe long-termhift in wind direction may be associated with acooling butarge area to become warmer. The effect ofand the complex interactions of air, water, and land makedifficult.
Climate Change in the USSR
etailed moieorological dm have been an Bible ftom the USSR dnce the. Moreover, dale for many ilaUocu eaiend beck to theh century. Summanei of these date, largely ibve, fotm the long-term rreraaes ased In tMi publication. Fredpttttlon and temperature data for tndMduaJ weather italkani In the USSR are broedcait two or fova timer dally by Sortti meteorological radio atatlom.ember of lha World Meteoroloelcal&pitlzatlon, the USSR iharei auch Information with foreign counulci. Theie data are put ol a. woitdwioe itanderdiied lyitem that rnwrtsconihtent quantitative measure! or weather perimeteri from year to year. There are nowtaUoni located In Ihe Soviet pain re-Hon. Theee data. when, aumnwtied, tern ea the base fot analytlng crop eondiOom. itudytng cUmate changes, eillnw'jna. Soviet piBi production, and evaluating th* affecu of changing technology on So MetS, average nluei of lempvteture, precipitation, and aoD molnuie for the pain hell were computed by hand; atnc* then they have been routinely procnied by compute:ally baiK The differences In proceulng could remit Inrandom etiore for Individual crop region* priorut countrywide annual average, wotddaff act ad by the method weed to calculate Un averages.
the past ISignificant fluctuation occurred in thethe Soviet grain belt that affected production. Examination of dailyprecipitation data foreporting stations scatteredSoviet grain belt suggests that while the Northern Hemisphere wasclimate in this area generally has become warmer andheIn the Soviet grain belt was causedhift in wind patterns.variations occurred among grain-growing regions from year to year, thethe USSR resulted from increasing flows of maritime air from the Northincreased precipitation and caused warmer winters and cooler summers32verage annual precipitation in thebelt increased byillimeters, or, in comparison with the
P.
4
long-term uvcragc priorhile average temperatures rose about tlircc-fourllisegreeoreover, the southward shift of the polar vortex during this period appears to have moved the moister northern climate southwardegrees of latitude, pushing back the desert and increasing agricultural production in the new londs area of Kazakhstan and Central Asia.
periods are used to discuss recent Soviet climate:
limate average of all available data prior0 is usedenchmark for comparison of subsequent events.*
hiseriod of transition from the dry, unstableo the moist, stable.
eriod of unusually moist, stable climate that was favorable to grain production.
This last period was unusually moist in the steppe regions of the grain belt.The cycle of alternate moist and dry years that is common in these regions was absent. The dryness5 combined with other global climate changes may indicate that the aberration of dependable moisture in this normally dry region has ended.
The choice of these periods was based on both climate events and on data limitations. Detailed data have been available only0 for precipitation and2 for temperature, marking the earliest dates that yearly analysis was possible. The break89 was chosen because of Soviet climate patterns and associated hemispheric climate events. The finalas separated because of its sharp departure from recent Soviet climate. These periods are used whenever possible In future discussion in this report. Lack of temperature data will at times01 to be omitted from the second period.
Changes In the number of crop regions falling Into the various climate typesough gauge of climate change. In thehe number of regions classified as either steppe or desert rangedoith an average of II (seen contrast,94 the average number of regions
i ctpfuito* prkveighted bym mm2
tffcneK* ofafaneicnt.nd.rd deliuoeipie tbe
No etNmpi ii made to edebibtt ihb periodliMst" Hi nlytimi itIftamry longerf record.
li
M
'
l^iti-tiiit
I V.
Dn
7 ShA lit*
ifi
htmIMM
Obit*
i.Mi.,
'
It*)
M
J
H
S
H
II
11
1
H
1
II
ft
ft
H
y
t
s
l
s
t
i
t
ft
$
ft
i
0
1
%
Ml mm
in
classified is steppe or desert has beenseehereharp departure from this recent trendS whenf theegions qualified as cither steppe or desert climates. Not since the drought3 had any crop region been classifiedesert nor had there beenr more crop regions that were too dry lo be classified as moist- continental. Thus, the climate5 would seem to fit better In thehan the.
comparing the climate0 with the long-term average,stand out. First,ew exceptions, thereteadythe climate of the Soviet grain belt0econd, thethendas more of an aberration than was thethe early and0 Viewed5 does notdryness when compared with the long-term average precipitation forbelt (seeather, the unusual climate in the USSR, as it did in many other parts of the world.
Effect of Climate Changes on Grain Production"
conditions in the USSR arc generally unfavorable to thegrain; both temperature and precipitation, acting alone and in combination,grain-growing region (seeorthorth latitude theare too cool and the growing season too short for grain.nd east of the Urals, Insufficient moisture prevents grain growingirrigation. The southern movement of the climate duringnd early
long-terra average number of rteppe tnd detent reglonihile the averagea*
winter and ipring grains are grown lit the USSR. Became of their differing need* they anIn different ami, but both typei of grain have tome require menu tn common. Winter grains,planted In October, lie dormant during the wtnter, tie hammed In June-August, have higher yieldsgxalni. and an uauaJty choein In regions whan both type* can be grown. Winter grain* do notlong lununct growing leaioo and can therefore be grown farther north than can taringintermoteled to the western half of tha grain bdt became winter tail ternparanret In the emit are toowtnter UH of the dormant pleats. (CrKVeal aofl temperature* at the depth of the tile ring modeTor winternd fot winter
Spring grainsifferent problem, ibtce almost all of them are groans In moliture-dcncJeni regloni Mort imall grelni grow bett when ioil mow lure it approximately half theatlmeten -and yleldi arlD be reduced If more or leuresent. Although excess mcliture frequently affecti winter gralru tn the extreme northweil. It li unuiual In the ipring grain regions where Increaacd precipitation generally meani Increased yleldi, and dxoughl ii the major problem. Both long periodi of low ioB moliture and lukhowyi (hot. dry eiiterly or southaatteriy adndiollomeieTi per hour and relative humldltlei> take their toD In various amounu each yew. Damage tosually quick and Irreparable, but Its magnitude dapendi on crop variety, phase of development, toil moliture, and lukhovey Intensity.
1 I
ubstantially boosted Soviet grain production, particularly in the spring wheat area along the southern fringes of the grain belt east of the Urals.
ecause or III Importance, wheat yield* and production data more comUtcnt and more complete than ere data for other grains.
o measure the impact of climate change on grain production, we have used spring and winter wheat dataurrogate for all grain. This avoids data problems arising from changing shares in sown area over time among grains with different yields. Wheat is the most Important rood grain in the Soviet Union, constitutingf the food grain production and about half of total grainoreover, wheat is grown bothinter and spring grain in the same regions as other important winter and spring grains, and it reacts to the
weather in the same way that other grainssing wheat yields as an indicatorf the variance of all winter grain yieldsf all the spring grain yields to be accounted for. In general in the Soviet Union, statements about wheat yields can be extended to other grains grown during the same seasons.
U. Winter palm Include winter wheel, rye, and aome beiley. Spring palm Include wheel, rice, oati, and moil of the barley. Com It normally referred toummer crof because ofong powtnj mice and late maturationffected by weather In much the aame manner as ipring pslni.
Seaor icurce* and methodology of these models.
Technology Include, application oferbktdei. and InKctfcWei; Implement, In teed nrieiwt: use of machinery: and application or aponomkailyprinciple, of crop rotation, fallow, and toll manage mem.
ethcae tool eonmnt.eei th. technologyd-ance. are not avert, applied to the entire country and each region iho wed duTcrent technology trends.
he impact of the climate change2 on grain yields was estimated by models constructed to calculate winter and spring wheat yields using weatherverage monthly temperatures, precipation, and soilnd multivariate regressionhe variance In grain yields is assumed to come from two sources: weather andocal differences in soil types are assumed to have no bearing on changes in yields overhe models explain
f the voriance of winter wheat yieldsf the variance of spring wheat yields. The remaining variance is assumed to be due to random errors in the data and short duration weather events not reflected in the monthly averages. Neither of these alter the weather trends.
ariations in the weathei caused production to fluctuate sharply from year to year, rangingowillion tons3ighillion tonsevertheless,24 total grain production in the Soviet grain belt increased at an averageillion tons per year (see'7 This increase resulted from technological improvementslimate change that generally favored grain production. According to the models, over half of this increase is directly attributable to improved climate,18
Table 3
USSR- Average Annual Change In All Grain
Increase In
by
Area1
Grain Production
Change
Hectares)
Metric Tons/Year)
Metric Tons/Year)
EtfLmated grain area ii based onaverage harvested area during lb* mostean foe which data .re available. Therop regloni used in lha modeb accounlillion hectare* ofillion hectares usually devoted lo grain.
The Soviet Grain Belt refers to therop regloni used In the computer model. These regions accountf tha sown area and moref the total production of ill pains in the Soviet Union.
These estimates have been normalized to remove variations In total sown area and relative sown areas of various cropa within crop regloni. Changes In area are not as pronounced2 as they were during the opening of the virgin lands in Kazakhstan In, and their effect on the results of this experiment can be ignored. Difficulties tn measuring the effect of weather (Appendix B) have already Introduced lomo uncertainty In the relative offsets or technology and weather, and the small adjustments for fluctuations inarea would only make the talk more complicated without impro>lng the accuracy of the result
he Impact of climate change was neither smooth nor steady during the period, nor did the impact occur equally in all regions. It is impossible to isolate
of spring wheat yields can be explained by climate change. Yields estimated by tho model without considering time trends show climate-related trends identical to the total trend or yields. The mean monthly precipitation from October through Septemberrend of the same magnitude as that ofthe spring wheat yields. Spring wheat yields rose by an average4 centners per hectare persec Figuref the Increase in spring wheat yields was due to the improved climate. Because most of the improved climate conditions occurred east of the Urals, spring wheat yields gained disproportionately.
reduction of variability of summer precipitation in the springalso has occurredistorically, spring wheat yields, whichprecipitation, haveuasi-biennial cycle. This approximate two-yearalternate wet and dry years exhibits periods of relative quiet (little oreveryo26 the spring wheatand halved on alternate years,eriod of highear-to-year changes were on the order, markinga stable period. The downturn4 followed by the sharp drop Inmark the start of another highly variable period.
Prospects
he exuttneeuiiHknnUJ cycla InSoviel Union hai been dlieuiied by many Soviet meieofologiiii. ctlrnalologbtt, and esTonombti For. Kolcikov. In hiiUnsnr Flew bi Agrtn,Hwt rnrnt AgrocD/mitt Zomtng,tacuiaes the poulble mo of thh featureoreoaitlni tool forether. Thh cycla Ii vary peniitaat. bul It tarue biennialery pronouncederiod of yean; then II may not be evident for wveral yean. In (be grain region of the Soviet Union ihb cyde of large and imall biennial fluctuation! requlrei aboutean to complete, Thb does not. however, mean that weather repeat! iudf everyean, bul rather that the wide fluctuation! ofnd the itable weather of theormal occurence.nusual ta the high precipitation tewj at which ihu liable period occurred.
Kctosko.ossible explanation for the qeaMkeuUel cyde. whkh may ot may Dot oa accurate. Simply Mated, warm winter* are followed, moit probably, by warm aprlngs. Thhind now that brlnai rainfall and clou da lhal In turnool eumnm. The cool lummer leti up condition that leadold winter, whkh glvai ihe reveci* weather th* followingot. dry lummet then kadi to warm winter and the cyde bsgiiu over again.
Soviet Tenth Five-Yearet grain productionanillionear during the period, about one-fifthwas realized on average. Attainment of this goal willthe most part, on the climate. Theoretically, the future climate mayimprove, or worsen in terms of impact on grain growing. Inthe climate cannot be forecasted, reasonable limits can be set that inbe used as the basis to project grain production. Moreover, some climate trends
seem more Likely than others. For example, the trend04 toward grcate' precipitation is unlikely to continue because of the physical limitations of the atmosphere to transport water from the North Atlantic to the Soviet grain belt. The recent wet period in the Soviet grain belt was unusually wet when compared to the past and the worldwide climate conditions in the Northern Hemisphere that accompanied it seem to havet therefore seems more likely that the future course of the climate in the Soviet grain belt will be toward less precipitation, either gradually or rapidly.
During thishe entire Northern Hemisphere was markediinutJc fluctuation. The Sahetian drought end failures of the Indian mortioon appear to be linked to the increased rainfall tn the Sovtel Union. Though the causa of the fluctuation are unknown. Its global extent Is evident. Changes In hemispheric climate of this msgnltude Involve the movement of tremendous amounts of energy, which afect the temperature of land surfaces and oceans foryears and In turn influence future dictate. The recent ending of the drought in Africa and the increased rainfall from the Indian monsoons support the hypothesis that this dimate fluctuation has at hut temporarily ended
Data art usually assumed to be Independent, and the year being forecasted for li assumed to be part of tha same population that the itatLiUci were derived from.
TTte projected value5 of Sovietgrain yields basedrno trendsentners per hectare. The standard deviation of the differences between the time trend valuea and the. actual values4 centners pet hectare,1 standard deviations from the mean asimitne trot test givesprobabilityor values outside the4 centners per hectare.
Below we set forth three projections of yields for spring wheat and winter wheat. Two of the projections arc based on differing average climatic conditions for the Soviet grain belt over four-year periodshe average "worst years" and average "besthe remaining projection is based on the average conditions. The projections cannot be used to predict actual grain production. but they do indicate the likely limits to which production will occur. Moreover, the projectionsseful analytical tool in assessing the USSR's likelihood of attaining the Tenth Five-Yepr Plan grain goal.
Even the limits set by the "best" and "worst" projections must be viewed with caution. First, they represent four-year averages, and Soviet grain production is noteworthy for its sharp year-to-year deviations from the trend. For example, grain production5 illustrated the extremes to which year-to-year fluctuations occur; production was roughly two-thirds of the output
Stated alternatively, when statistics are used to forecast the future, certainre made that may not be true, having the effect of making the forecast meaningless.ime trend of Soviet all grain yields04 it can behat there ishance that5
yield would be higher4 centners per hectare or lower. The actual yieldhich Is well outside the widest range statistics can set as possible. Moreover, ifyear cycle of fluctuation in the amplitude of the quasi-biennial cycle has ended, as seemseriod of greater variability In grain yields can be expected.
Finally, the method used in arriving at these projections uses two assumptions that tend to inflate the lower estimates. First, the effect of the time trend of technology is assumed to be Independent of weather. The effect of weather on technology, particularly fertilizer application, tends to Increase variability of production by Increasing yields only when sufficient moisture is available. In dry years the application of mineral fertilizer may actually reduce production. The net result of this effect is that weather is nonlinear and therefore cannot be averaged before yields arc computed. If weathereriod of fluctuating moisture supply is used to compute annual yields that are then averaged, the resultower mean yield than if the weather is first averaged and then used toean yield directly. The ratio of the variance to the mean of yields25 is twice that of yields0
Second, the assumption that the "worst" weather possible occurred25 seems conservative. The worst four consecutive years of theean precipitation that is onlyelow the long-term average. The indication is that theay be more representative of average future weather than iteasonable bottom limit.
y combining past climate and an assumed constant technological development trend, future grain production can be estimated. The three projections were each based on mean weather parameters averaged over some past period (sechese estimates of average weather were then used to compute yields of winter and spring wheat in each crop region, appropriate technological trend adjustments were made, and all grain productions were computed for each yearor each period's average weather.
Protection I; The "Beit" Case
This projection is based on the climate. Except for the winter grain crophiseriod of stable, favorable weather in the grain belt Precipitation was abundant and temperatures moderate. Winters in the winter grain region were warm (with the exceptionnd summers In the spring grain region were cool. If
these climate conditions prevail, winter and spring wheat yields will09 centners per hectare. Grain production wouldillion tons per yearore than one-third greater than average production. This projection seems highly unlikely.
t>j* imt
its*
USSR: Spring Wheat Yields and Projections Based on Weather from Various Years
Projectionerage
This projection is based on the climate. This periodomplete cycle of the variations in the spring grain quasi-biennial cycle from the maximum fluctuations of thehrough the stable. The high precipitation of theore than made up for the dry, making this period wetter than the long-term mean priorhis climate gives projected winter and spring wheat mean yields48 centners per hectare, respectively. All grain annual production wouldillion tons per year under this assumption. Although this represents aneduction from projection I, it is still somewhat optimistic.
USSR: Winter Wheat Yields and Projections Based on Weather from Various Years
f'mjcciirm J: ihe "Worst" of the Recent Vast
This projection is based on the climate. This period marks the worst four consecutive years of the pastears. The worst yearhen production of both winter and spring grains suffered, and no year during the periodood year for both types of grain. Thisry period for the entire grain belt, with hot summers in the spring grain regions and cold winters in the winter grain regions. Precipitation, when compared with the climate of projectionasor the winter grainsor spring grains, and the highest June through July temperature in the spring grain regions in the pastean washis period was also marked by maximum amplitude of the quasi-biennial cycle in the spring grain regions. This climate gives projected winter and spring wheat mean yields0entners per hectare, which, in turn, give an estimated all grain mean annual productionillion tons. Thiseduction of projectiony means of comparison, the Soviet goal of an average crop ofillion tons during the current five-year period seems merely to be an extension ofrend. If the disastrous harvest5 is included in the trendowever, the average cropsons. In any case, based only on trends in production, Soviet leaders can reasonably expect grain productiono exceed output. However, considering the climatic conditions necessary to achieve their goal, it' appears unlikely that the goal will be met.
he methodologies underlying Ihe estimate for the first four categories are explained In detail. The Sorki Grainhe methodology for the feed climate it bated on offrcMI Soviet pirns for livestock product output during the five-year period.
he estimated annual requirements for grains estimatedillion tons, based on projections of normal requirements for food, seed, industry, exports, andhe projected level of production fallsillion tons below the projected level of requirements. However, the level of probable imports does not necessarily equal the gap between production and requirements because, in the event of smaller crops, requirements likely would be reduced, as they were.
fffK'li BLANK^*AGE
APPENDIX B
METHODOLOGY FOR SEPARATING THE EFFECTS OF TECHNOLOGY AND CLIMATE ON SOVIET GRAIN YIELDS
Grain yieldsunction of climate and technology. In the past it has been generally assumed that climate remains constant and that any upward trend in yields is due to technology, with annual weather difference causing the variance about this smooth trend line.loser look at the data suggests that24 thereignificant improvemcnv in the climate of the Soviet grain belt that makes this assumption invalid. The problem is therefore to separate the effects of technology from those of weather during this period. To doomputer model was designed that would accurately describe past yields for spring wheat, winter wheat, and all grain forrop regions, using monthly average weather data and estimates of technology derived from the residuals. These estimates of technology, when converted to fertilizer response rates, compare well with published Soviet data at the republic level of aggregation.
Formulation of the Model
Technology
During the pastears Soviet investment in agriculture has steadily increased. New varieties of their major grain types have been developed; fertilizer, lime, and insecticides have been used in greater quantity andore scientific manner, and new and better equipment has been made available to plant, care for, and harvest grain. These improvements are recognized as components of the technology trend, but insufficient data are available to estimate the quantitative effect of the improvement. Information on such improvements is incomplete and, when available, represents national or republicot the individual crop regions required for the model.
Although the effect of technology on grain yields can not be directly measured, certain Important assumptions can be made about the form of relationship that are believed to accurately reflect actual conditions.
The rate of change in application of technology, if present, will not change greatly from year to year for any one crop region. This allows the useinear trend for technology:
1)
where
* lime in year* c- AT/attonrlant
The effect of technology (excluding irrigation) is weather dependent. For example, the application of one centner of fertilizerillimeters of annual precipitation will increase yield more than will the same centner of fertilizer withillimeters of annual precipitation. The effect of technology is. therefore, notirst approximation of the effect is:
2)
rather than
where
Y-yield
yield baaed on weather condition! alone
Technology has not been evenly applied to the entire grain belt. There hasendency to apply technology first in those areas where the potential return was the greatest. The regions where climate is more favorable should have greater technology trends (greater values of c) than those with poorer climates. This was found to be the case.
Weather
Both technology and climate affect grain yields. The problem is to quantify the influence of each. Since thereelative abundance of weather data, the use of regression analysis allows us to estimate the effect of weather on yields and thereby to treat the influence of technologyo determine the impact of weather on yields, the weather data for all years2 for all
regions in the Soviet grain belt were regressed against annual yields for these regionshis assumes that soil types and other differing physical characteristics between crop regions have no bearing on yields. While this oversimplification is not true, the importance of these differences on yields is minimal. To test this hypothesis, regions were grouped by annual mean precipitation, and the weather-yield coefficients did not change significantly,usthe independence of yield to geographic location.
Data
The weather data used in the model include observations of precipitation and temperature for each ofrop regions2 to the present. The observations arc averaged and entered monthly into the data base for each crop region. In the early years ofe transcribed by hand from weather maps, greatly limiting the number of stations that could be used.5 the data have been routinely processed by computer, thus enabling all reporting stations to be included. The change in the number of reporting stations was one of spacial density and not of shifting areas. The effect on monthly mean data should be negligible when aggregated to the crop region level.
The yield data cover all grain, spring wheat, and winter wheat for. Yields were calculated from published data for oblastsrop region, except for those few crop regions that coincide with the areas for which the USSR reports yields. Errors in the yield data can arise from inconsistencies in official Soviet statistics, the need to estimate unpublished data, and the ratio of sown area to aggregate yields. Although many inconsistencies appear in published Soviet sources, mostariance ofentner per hectare. Wheat yields for the RSFSR are usually reported only for spring and winter wheat combined. By using additional information, yields for spring and winter wheat could be estimated separately. The estimates that could be checked were quite accurate; hence, any distortion is believed to be small.
The Model
Basic Equations
The model uses regression analysis of linear and quadratic weather parameters tolobal qualitative equation relating yield to weather:
(equation 4)
where
Y*qualitative yield based only on weatheregression coefficients
value of temperature, preclpltatliu, ortoll moisture averaged for any Bomber of whole) during the crop year. The crop year it arts In October and extend* through September of the year of harveaL The equare of Iho value la also used when
region
t-year
To compute the regression coefficients, all pairs of yield and weather data available for allrop regions forears are used simultaneously without regard to location or time.
To estimate the level and rate of change of technology in each of therop regions, the global equationbove) is solved for each year in each crop region,atio of the. qualitative weather yield to the reported yield is calculated. These ratios are the technology indexes:
5)
where
actual technology Index vreported yield
weather based ykld eiUmate lcrop
These technology indexes are then time trended tomooth function that removes random variance and allows valueso be estimated for years for which no yield data are available and to project future values. Thus:
6)
where
alculated technologyear
b| and c, are conitanta delermfned by linear regressionith Ume (an equation 5)
To estimate Ihe combined Impact of weather and technology on yields, each value of Y* Is multiplied by the corresponding valuehis allows yields to be computed for any crop region for any year in which the weather can by measured or estimated, Including future years.
where
yfcUoocOon of Urn tad
. ) here
and I, Uirough a, an ootutuiti Parameters and Coefficients
An almost limitless number of combinations of weather parameters can be used in describing the changes in Soviet grain yields. For example, the weather data base contains only monthly values of temperature, precipitation, and soil moisture. These basic parameters can be averaged or summed over any combination of consecutive months, combined with each other using any algebraic form, changed to any functional form, or combinations of all of these. An extreme example might
where
X" new paruaetrr jVf-ntttn petcfclttUoon tBcaperatUHJ Oct-Apr
arge number of parameters increases ine model's ability to describe past yields but reduces the physiological correctness of the weather-yield relationships. Care must te taken that better description is not achieved for the wrong reasons. An example of this Is the strong relationship between all grain yields and the average January-February temperature, whichf the variance of all grain yields. The reason for this correlation is that temperature is an indicator of the state of the climate. The usual climatic conditions that leadarm winter also produce above-iiverage precipitation through the next summer. The winteras one of the warmest on record, but was unusual because precipitation was far below normal. The entire year was dry, which reduced yields
rather than increasing them as the winter temperature had indicated would happen. In any descriptive model all parameters must be evaluated for physiological correctness before using them. The major controlling factor for grain production in the Soviet Union is moisture. Available soil moisture when growth starts in the spring results from precipitation that fell from the late fall of the previous year until the start or growth. Once growth begins, temperature as well as precipitation becomes critical, particularly in the case of spring wheat. The model's parameters and their coefficients are given innd the model's estimated yields are compared with the officially reported yields (seeo officially reported yield data are available5 at the crop region level.
USSR: Weather Pararneten and Coefficients for the Global Equation
All grain
Error
precipitation Nov-Apr'
precipitation
above squared
wheal
precipitation Nov-Apr
precipitation May-Jun
above squared
ooi ii
wheat
temperature Jun
temperature Jul
precipitation May-Jun.
soQ moisture Jun-JiJ1
above squired
monthly precipitation over the months lodicatrd.
Soil moisture li calculated from temperature and precipitation daBy data and reported rnorrthty, Vahsei rangeiUlmeser*.
Measuring Technology
Substitution of weather data into the global equation explainsf the variance of the yieldsdding the means (equationnd linear time trends for each crop region for technology indexes increases the
USSRot KrponHt
c
porud
0
6
4
I*
variance, depending on the crop. The remaining variance Is due to random noise In the data caused in partack of resolution.
To separate the effects of technology and weather, the yields2 for each crop region for each year were adjusted2 technology (see
USSR:Reported AO Cob., AdtuHd1 TnWoo
trt Heel tit
] 4 S 6 7 8 9 0 1 2 3 4
R* potted
tchnototr BtkmiHlt
Ra ported
Ukraine
Reported
oldivla
Re ported
Nhnolocy RSFSR
Reported
IMltnrtnoletjr Kt/Mhttin
Reported
KhMlofy
6
8.2
44
4.0
M
IU 74
1 5 4 2 7
5
7.1
7*
*
J
138
104
84
7.9 8J
61
it
M BJ
By doing (his, variutions in yield ore caused solely by weather. The adjusted yields were computed using the following equation':
(equalion9)
where
ield2eported- Intercept andfrom" crop region
It is not possible to explain all the weather effect using the data base available, and therefore It Is not possible to prove from the weather and yield data alone that this is inuantitative measure of technology. However, Soviet reports of fertilizer applications (the main technological development) by republic are consistent with the model's measurements.
In formulating the model, yield wii luimcd to be due to the combined eflecti uf weather end technology. An animate of the wither effect; In Ihe formield from the global equation, wii multiplied by the appropriate lecruology Index to get the ertlmiied yield, To adjuit any reported yield to the technologyhe proeeeirevened. The reported yield ii divided by the appropriate year'j technology Index to get the aiaumed weather effect, which la then multiplied by2 technology rndex toield2 technology.
equation fl
equation 6*
laerefore
Udng the total time trend ai afor technologyresponse ratei that vary In contrait.n Bdonmla5 In Katakhrten.
Fertilizer response rates are computed from the current nutrient application rate and the yearly changes in yield, assuming that application rates were zero2 (seesing the calculated technology trends results in response rates that are very close to the rate reported by Soviet
USSR: Trend InertUIier Application Rates, and Response; Rates
I. He <ponsc rates are computed assuming (t>f thegron weight of app Med (eitllticr ii nuttlenthat no fori Mm" was applied as2inear increase occurred each yearoviet data Indicateirtually no fertilizer wis applied2lthough Increases in applications have been greater In recent years than In earlier yearsonlinearinear npptoximnli quite cbte to the actual application rales over this period.
- reiponso rate
average annual change In yield (centners per hectare per4 nutrient application tale (centners perears tratio of nutrients to grow weight of Average annual,
Average Annual Increase In All GrainCentners per Hectare)
Rate1
to
to
to
Increase
her and
4
Yield per Unli
per Hectare)
of Fertilizer)
Original document.
Comment about this article, ask questions, or add new information about this topic: