f C P ' Yielc' Mod?! Development

i,

& ) I ! YM-N4-04456 I < ."

-=:

JSC 18908 -s

k;: $4

IS YO^ . 2 , t

A Joint Program for -. $ rfg sg

, "i Agriculture and , b . -;

. . Resources Inventory .,.( > .

Surveys Through -3 % J

Aerospace , ? *P

Remote Sensing

JANUARY 16, 1934

.* . (E84- 10 105) A B G l i N T I N A CCBN Y I E L C HODEL b184-2 1923

-2 ; - . (National Cceauic and Atmos~heric

Bdrinis trat icn) I S p riC A02/!lF A01 C S C L 02C : 1 * 5

Uilcla s y i * a +

G3/43 00105 h !

1 SUSAN L, C A L L I S 5 CLARENCE SAKAMOTO

i USDA/NOAA 1 C IAQ/RODELF %RANCH

ROO1 I 2 C) E ERA ELDG 608 cHFRRY STRE$ COLUMBIA, MO 6 201

Lyndon B. Johnson Space Center Houston Texas 77058

1

https://ntrs.nasa.gov/search.jsp?R=19840013855 2020-06-27T00:20:27+00:00Z

ARGENTIKA CORX YIELD ?lODEL

Susan L. Callis

and

Clarence Sakamoto

A I S C Models Branch

January 16, 1984

1. R@mrt No. 2. -t Atcamion No. 3. hapcant's Cltalog No. YM-N4-04456, JSC 18908

Argentina Corn Y ie ld Model I 6. M o r m ~ n g Organ~zat~a 1 I

Susan L. Cal l i s and C l arence Sa kamoto 10. W a k Unit No.

9. Paforming Oqmiotion Nwm and Addms

USDCINOAA C IADIModel s Branch Room 200, Federal Bldg, 608 Cherry S t . Columbia, MO 65201

b 16. Abrcrat

A model based on mu l t i p l e regression was developed t o estimate corn y i e l d s f o r the country o f Argentina. A meteorological data set was obtained f o r the country by averaging data f o r s ta t ions w i t h i n the corn-growing area. Predictor var iab les

I I 19. Srwitv Clamif. (of this rrportl 1 10. SIcurity Classif. (of this prprl 1 21. NO. of pages 1 22. Rice* t

i

11. Contract or Grant NO. f

13. T v ~ of ReW't and Puled Covrrd

i i 1

1

12. slmnuwl~ A W ~ ~ V N8llm Adbat National Aeronautics and Space Administrat ion Lyndon B. Johnson Space Center 14. ~oonwlng ~gency c06 Houston, TX 77058

17. Key Words ( S u w bv AuMw(sl)

Mu1 t i p l e regression Predictor var iables

*For rJI by thr Nationrl Technical Information Srvico. Springfield, Virginia 22161

JSC Form 1424 (Rw M a 78)

,

18. Distribut~on Statement

I Unclassi f ied

NASA - S C I

f o r the model uere derived from monthly t o t a l p rec ip i t a t i on , average monthly mean temperature, and average monthly maximum temperature. A " t rend var iab le " was included f o r the years 1965 t o 1980 since an increasing trend i n y i e l ds due t o

Unclass i f ied 1 13 I

1 technology was observed between these years. t

ORIG1MAL PAGE IS OF POOR QUALm

I NTRODUCT I ON

The purDose of tn?s stuay was t o se lec t monthly weatner var laa les t ha t coula I C *

De used t o estimate y i e l d s of the Argentina corn crop. Tpe corn-grow~ng area

i n Argentina i s ind icated i n Figure 1. Most o f Aryent lna's corn i s grown i n a

concentrated area i n cent ra l Argentina where t he c l imate i s predominantly humid I subtropical . The western edge of t he soybean area i n cen t ra l Cordoba i s semi-

a r i d w i t h warm t o hot summers. There drought and h igh temperatures can be a

problem dur ing t he growing season which begins w i t h p l a n t i n g i n October and

November and extends t o harvest i n A p r i l through June.

METHOD

M u l t i p l e regression analys is o f y i e l d w i t h selected agrocl imat i c ind ices

was used t o der i ve a su i tab le model. The regression equation f o r t he corn model I der ived i s :

where PI Y = Es t ima tedy ie l d

a = Constant

B j Coe f f i c ien ts o f the var iab les j = 1 - 3

T = Trend 1965-1980 A A - -

ETMETHi = ET - ET f o r month i: where E T = K * P E T and K = ET/PET.

T X ~ = Maximum temperature f o r month i, and

E = Unexplained er ror . \

Trend between the years 1965 and 1980 was determined from a p l o t o f y i e l d s as

shown i n Figure 2. ET (evapotranspirat ion) minus & ( c l i m a t i c a l l y appropr iate I evapotranspirat ion) was used as an index represent ing so i 1 moisture fo r p l an t 1 : growth. A more complete d e f i n i t i o n i s given i n t he Appendix. Large p o s i t i v e

A E T - ET values suggest wet condi t ions.

- - . . - -- - - - , - - - - -. -- - -- - - - - , , - -&--&--ma;- m :*--.- , ; , , - l*&*..i *-=----- - - ' , - ( . - _

i m

,., tmr-maul bun- - --- ~ " , , , u l & o o C q I - P w i b C y

p- C ~ r r

b p l e Wac ~CPtat

Figure 1 . DENSITY OF P U V T E D AREA IN CORS 1977178 CROP YEAR (Source: "Agronomic Characcerimtlon o f the Arren- cina Indicator Regiun")

ORlGlNIL PAGG f3 OF POOR QUALITY

v 1 6 I I I I I , I . , . . . , I m . . , , , ,

Y E A R 60 6s 7 0

CORN PRODUCTION

Figure 2. Plots of Production and Yield Versus Year for Argentina Corn.

~ -

CORN YLtLDS

x" :~ '~ - - . - - - - z ,, , . '..,.. , .

10 developing t he models, var ious procedures o f t he S t a t i s t i c a l Analysis

System ! S A S I n s t i t u t e , Inc., !o?9! were used. The proceCures used and the

.. L,,,,, -a* C , I O ? ~ - - ?errf3~rted ~ f t n eacn are sumnari zec i n tne L.?;lcncrx. Tne selectec

;nose; :;ic trle R: g w s t 22 ana inciuded v a r i ~ b i e s tha; were s i g r l i f i can t a t t n e

10 per cent l e v e l and were agronomically meaningful.

DATA

Crop data f o r Argentina from 1960 t o 1979 was obtained from the Foreign

Ag r i cu l t u ra l Service (A1 an Vandagriih, personal communication, 1982). A data

set was created w i t h year o f y i e l d as year o f p lant ing. Since the grawing Y d

season spans two numerical years, t he crop da ta was expressed i n terns o f "year

+ 1" (o r harvested year) and the meteoro log ica l data corresponding t o any y i e l d

included data f o r t h a t year and "lagged" da ta f o r the previous year.

The meteor01 og ica l data was created us ing t h e general Argentina

meteorological s t a t i o n f i l e . This f i l e was composed o f data from several

d i f f e r e n t sources, i nc l ud ing t he Monthly C l ima t i c Data f o r t he World and t h e

Se rv i c i o Meteorological Nacional i n Argent ina (R.E. Jensen, C.M. Sakamoto, and

S.E. Mummert; August, 1974). S ta t ions i n s i d e t h e corn-producing area were

averaged by prov ince and weighted according t o t h e percentage t h a t t h e i r pro-

v ince con t r ibu ted t o the e n t i r e count ry 's product ion. F igure 3 shows t h e corn-

growing area i n Argentina and associated s ta t ions ; Table 1 l i s t s t he s t a t i ons - used and t h e i r weights.

* The years between 1961 and 1980 (year o f harvest ) were used i n t he model

s ince these years contained t he most complete meteorological data.

PROCEDURES

Weather va r iab les were selected from c o r r e l a t i o n s and p l o t s w i t h y i e l d .

These var iab les, along w i t h TREND, were t r i e d i n regress ion equations. The

1 i ln i ted number of data years severely l i m i t s t he nunber o f var iab les t h a t

.--. ] . - -*.- - - -"a'-.. vr-

ORIGINAL PAGE IS OF POOR QUALlW

C A

b

B e l l V I I ie l . Rio C u r r t o

Trenque Lauquen

i BUENOS AIRES

M8e8ehln l

-

J

BUENOS AIRES

oRlOMAL ? A M IS OF POOR QUALW

Pe rgamino Junin Nueve de Julio

CORDOBA

Rio Cuarto Bel l Vi l le

SANTA FE

Rosario Casilda Las Delicias Parana Pe rgamino

Table 1. Meteorological Stations Used for the Argentina Corn Model.

ORIGINAL PAGE 121 OF POOR QUALITY

can be included i n a regression model by l i m i t i n g t h e number o f degrees o f

freedom. This i ~ f l uenced the choice o f model considered "best" f o r soybeans.

Tnree weather var iab les plus THEND were considered the optimum moael w i t h 19

years o f data. Tne tnree weatner va r iab les are: A

ETMETH11.. ......... ...ET . ET f o r November

TX12 ............... ...Maximum temperature f o r December

Txl................... Maximum temperature f o r January.

The c o e f f i c i e n t f o r t he November va r i ab l e was p o s i t i v e which r e f l e c t s

moisture requirements j u s t a f t e r p lant ing. The c o e f f i c i e n t s f o r maximum

temperature f o r Decenber and January were negat ive i n d i c a t i n g t h a t temperatures

t oo h igh i n t he peak summer months du r i ng t h e t a s s e l l i n g and s i l k i n g stages lead

t o s t ress on t he plant. There i s a c r i t i c a l growth per iod f o r corn. The

s t a t i s t i c s o f the selected model are summarized i n Table 2.

TEST RESULTS

A jackkn i fe t e s t was run on t he model. I n t h i s t e s t a year was e l im ina ted

from the crop data and t he model was used t o p r e d i c t t h a t year ' s y i e l d . This

process was dvv . f o r each successive year beginning w i t h 1961. The r e s u l t s were

reasonable w i t h the greatest d i f fe rence (3.24 qu in t 31 s) between p red ic ted and

actua l y i e l d i n 1976. That year severe drought badly damaged the corn crop and

harvest was hampered by ra ins. Test r e s u l t s a re p r i n t ed on Table 3 and p l o t t e d

on Figure 4.

. . -- (I, *--

APPEND I X

De f in i t l on o f Variables

Tne d i f iorence between E i and E f 1 s an index o f tne amount o f moisture

avai l ab le fo r p lant growth. Soi 1 moisture aepiet ion i s base0 on evapo-

t ransp i ra t ion (ET) estimates, determined as f o l lows:

where

(ET), = "Actual" evapotranspiration,

( 1 = Available moisture a t end o f n-1 month,

AWC = Maximum water holding capacity,

(P), = Prec ip i ta t ion f o r month n, and

(PET), = Potenti a1 evapotranspi ra"ion f o r month n. A

ET - E T measures the d i f ference between the actual evapotranspiration and the

" c l imat ica l l y appropriate" evapotranspiration, g i v ing an i ndicat ion o f s o i l

moisture supply and demand.

S t a t i s t i c a l Analysis System Procedures Used

PROC CORR

PROC PLOT

PROC STEPW I SE

PROC STEPW I SE FORWARD

Ccnputes co r re la t i on coe f f i c i en ts between variables, i ncl udi ng Pearson product-moment and weighted product-moment correlat ion.

Graphs one var iable against another, jroducing a p r i n te r p l o t . Provides f i v e methods f o r stepwise regression. Stepwi se i s useful when select ing variab,es t o be included i n a regression model from a co l l ec t i on o f independent varlables.

Begins by f i nd ing the one-variable model t ha t produces the highest ~ 2 . For each o f the other i ndependent variables, FORWARD calculates F- s t a t i s t i c s r e f l e c t i n g the contr ibut ion t o the model i f the var iab le were t o be :ncluded.

PROC STEPWISE MAXR

PROC PETM

PROC ZINDEX

ORIGINAL ?Ad€ 1(1 OF POOR QUALIW

Begins by c a l c u l a t i n g s t a t i s t i c s f o r a model i nc l ud ing a1 1 t he f nde9endent var iab les. The var iab les are deletea f ro? cne model one by one un:il a l l t n c r25z:r . ; r ; v a r i a ~ l e s produce i - s t a t i s t i c s s i g n i + t c a n r 3: : le .lil i eve l .

Tne stepvJisr metnod i s s moe: f~cat ior : o f t n r forwara se lec t ton tecnnique, d l f f e r i n g i n t h a t va r iab les already i n t h e model do no t necessar i ly stay there. A f t e r a va r i ab l e i s added (as i n t he forward se lec t i on method) t he stepwise method looks a t a l l t he va r iab les a1 ready included i n t h e model and de le tes any va r i ab l e t h a t does not produce an F - s t a t i s t i c s i g n i f i c a n t a t t he .10 leve l . Only a f t e r t h l s

*

check i s made and t h e necessary de le t i ons accomplished can another va r i ab l e be added t o t h e model.

(Yaximum ~2 improvement) Un l i ke t h e th ree techniques above, t h i s method does no t s e t t l e on a s i ng le method. Instead i t looks f o r t h e "best" two-var iab le model, t he "best8' t h ree va r i ab l e model, and so f o r t h .

Uses l a t i t u d e and mean monthly temperature t o ca l cu l a te Thornthwai t e ' s po ten t i a l evapo- t r a n s p i r a t i o n f o r each month.

Uses monthly PET'S, p r t . i p i t a t i o n , SS (beginning mois ture i n surface layer ) , AWCS ( ava i l ab l e water capac i ty i n surface 1 ayer, SU (beginn ing mois ture i n t he under ly ing 1 ayer) , and AWCU (avai 1 ab le water capac i ty i n t he under l y ing l aye r ) t o c a l c u l a t e

mois ture budget, drought index

LC 'r S f , P fi\*-&b*%.r&- t ze%=:us#xr -522 it' r2?,3*?+?*>***2 - --***%9e* J F - = * ~ OWQHAL PA= 19 ............... k i3 if'.?\,^*

*?-*-9*--!-#-9*-- OF POOR QUALITY

U t I t l l 1 l I l l l t t t t

ORIGINAL PAGE 13 OF POOR QUALITY

I 6 *- I P t

Q

(It