General Disclaimer One Or More of the Following Statements May Affect

General Disclaimer

One or more of the Following Statements may affect this Document

This document has been reproduced from the best copy furnished by the organizational source. It is being released in the interest of making available as much information as possible.

This document may contain data, which exceeds the sheet parameters. It was furnished in this condition by the organizational source and is the best copy available.

This document may contain tone-on-tone or color graphs, charts and/or pictures, which have been reproduced in black and white.

This document is paginated as submitted by the original source.

Portions of this document are not fully legible due to the historical nature of some of the material. However, it is the best reproduction available from the original submission.

Produced by the NASA Center for Aerospace Information (CASI) 'lade available u n ler y4SA If, Me Icterest al e3 1 :y anti ,;?e ,fi3- j s! r.-laCen a Earth Resoul:es SJ*vey 7, 8— 1 01 8 OL 1 information 3uG w+Uloul '73 101 any Use RI:i k thrri:o ` A `p -' ^!

(F78-101P9) rJSE CF LPNCS47 Il!zCHY FCF SCIL N78-3148F SURVEY (Institutc de Fesquisas EsFaciais, Sao Jose) 24 p HC 402 /hF 401 CSCL OPtt Pnclas G3/43 00185 KtCEIVED BY SA STI FACIL — DATE: _ ZZ - DCAr ^I v Z c^ EJ BY C I' ^: + S 'AGILITY EaA - SUS AIAA

/ 0^^(1111 ^ 1^^^

u t .' '^IwSTiTUICONSF.LNO NACIONAL u DEDE DESENVOLVIMENTOPESQUISAS CIENT(FICO ESPACIAI8^^^E TECNOLO CP ^- 11

1 M INDEX

MAP1ER I - INTRODUCTION ...... 1

CHAPTER II - MATERIAL AND METHODS ...... 3 CHAPTER III - RESULTS AND DISCUSSION ...... 7 3.1 - Climatic Conditions ...... 7 3 .2 - Land Use ...... 7 3.3 - Natural Vegetation ...... 7 3 .4 - Relief ...... 12 3.5 - Drainage ...... 12 3.6 - Spectral Measurements ...... 17 CHAPTER IV - CONCLUSIONS ...... 23 BIBLIOGRAPHY ...... 24

G RAGE RAW MT

- iii - PUMDM PAGR NOr ., CHAPTER I

INTRODUCTION

Soil survey is one of the main methods to attain rational land utilization.

Mang the difficulties encountered in tarring out traditional soil surveys are the high cost involved and long time required for the execution. In Brazil, these difficulties are aggravated by the size of the country and by the difficulties to access some areas.

New technologies and methods have been developed to minimize these problems. Now,imagery from the LANDSAT series of satellites allow a global, repetitive and multispectral coverage of the whole Brazil- ian territory. The pr•pose of this work is to evaluate the LANDSAT system capability for soil survey.

This evaluation was performed based on the comparison of the results obtained by visual interpretation of images, taking into consideration the following: land use, natural vegetation, relief, drainage pattern and :.pectral response, as well as comparison with results of other researchers who had used aerial photographs and existing soil maps. _ T

CHAPTER II

MATERIAL AND METHODS

The selected area - Ribeirao Preto - is located in the NE of the state of Sao Paulo and the SE of the state of Minas Gerais (Fig. II.1). It lies between the latitudes 20 030' and 22030', South and the longitudes 47000' and 49000' West.

To accomplish this work, LANDSAT images(four bands)in the scales of 1:1,000,000 and 1:500,000 were used as well as transparencies in the scale of 1:3,369,000 and 1:1,000,000. The MSS images were acquired on July 21, and August 18, 1973.

Color compositions done in the Color Additive Viewer - I 2S were utilized to aid in the interpretation. The transparencies in the 1:1,000,000 scale were used to quantify different hues, related to different soil groups, which were analyzed in the Image-100 System.

Spectral measurements for the various groups of soils were obtained with the Earth Ground Truth Radiometer. Spectral response curves were also detel-mined in the Laboratory using a spectrophotometer operating in the visi;,le range.

A study of the water balance for Ribeirao Preto, according to Thornthwaite's(1948) method was done to evaluate the soil capacity.

Visual interpretation was carried out in black and white images at the scale of 1:1,000,000, (channels 4,5,6 and 7) to map the following themes: land use, natural vegetation, superficial drainage and relief. After this initial mapping,these themes were also mapped on the 1:500,000 images.

Different categories of land use were characterized on the MSS images based on texture and spectral responses with the aid of color compositions. Based on available bibliographic information and sketches of land use, it was possible to infer the spatial distribution PAGR MANg NOT FnW.

-4- m ,

IttlGINAL PAGE IS F P(X)R QUALITY -5-

of different types of vegetation likely to exist or having existed in a certain area.

Qualitative and quantitative studies of the drainage patterns obtained by manual interpretation were done to evaluate the possibility of using drainage pattern for soil classification.

The qualitative study was developed according to Lueder (1959). The quantitative factors analyzed were: drainage density (Ray and Fischer 1960), stream frequency and texture ratio in circular samples (Souza, 1975). These calculations were based on circular samples with an area of 110 kW. For each homogeneous drainage area, the average indices values were submitted to a Correlation and Regression Analysis.

The final results were compared to those found by other researchers who had used aerial photographs and conventional soil maps. CHAPTER III

RESULTS AND DISCUSSION

3.1 - CLIMATIC CONDITIONS

The Ribeirao Preto climatic study was based on meteorologic

data obtained as monthly averages over several wears prior to 1973 and more detailed data for 1973.

The conclusions are summarized in Figure III.1. Inferences on natural vegetation and soils were made based on water content. It has

been observed that the spectral response of the soil and the vegetation changes with the water content of the soil.

3.2 - LAND USE

Channels 5 and 7 proved to be better for distinguishing

land use categories. Visual interpretation of the images permitted the classification of land use into 6 classes: crop areas and grassland, "cameo cerrado", natural forest, gallery forest, artificial forest and urban areas. (Fig. III.2).

Comparing a conventional soil map shown in Fig. III.3 with the land use map obtained from LANDSAT images (Fig. III.2), it was possible to identify soils in good agricultural conditions like Latossols.

Observations of land use were helpful to infer the soil groups in the area. These observations agree with Elbersen's (1973), Cipra's (1973)

and Hilwig's (1974) results.

3.3 - NATURAL VEGETATION

A map of natural vegetation was prepared based on information gathered from a conventional soil map (Buckmans and Brady, 1968), climatic maps, and remaining natural vegetation (Fig. III.4).

RRBQCDMG PAGE MANX NOT ALMI.

goo Prstife - t^3s sun Pat. Trwo - 1092 e

I - 356 ms, tMai1 iWiet. - 4 nnn

too

C n

0 Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sep. Oct. Nov. Dew.

Surplus Moisture. - D.,pletion Moisture.

- Deficit Moisture. - Recharge Moisture.

Fig. III.1 - Water balance for the "Ribeirao Preto" region.

ORIGINAL PAGE IS OF POOR QUALITY

J ^ ^. 4 •4 t

4.^ t .F3 s a

q ,^,

Cry.} J/' , a ^ ..,t r ^ r.: C

N11 v %S1' t ri ._ l^t ,. ^^ „^ - e ) ^ l )1+" ^^.. J.f^ - r't irttlS ^+„yt •jam F . ^n ~ J L E [ • ^`1

`tiel ^. :.1 'l..Jr+^-itJ ^ ^`^^'i^_y"^t`..^ll , ^ .= stt .-^ •^Jt^ ^.f,w/w ^ ^F d :^,ps t^'It ,.^ L 'd fy1 t - - ^" ^4. Fe.^^t '^ ,! - S" _ y.,•^^,: _ ^"'fit. .^S yli ,•'/^h1'f ... tom - `i. ^^...,"'-^ri";^^4^`.+ ^ !^y 1..!i `^ /^A,I^rY ' r! 7 cn e r yt/ } ! y , lt1/•. e, ';`F^ (r G^ i • .r^~ r^ ^ C . i- ^^ t'?}^s•,-t^ .r ` rt ^ ^•,^J^'•:. Y•^^'G d^-•./.^ _ s` e t ^ ^/rl}(.^ j •w' ,^y.' . ^ S .^ ^^,; ^ t^^ 1 ^ ^ e ^ Q D D ti

^ • 7'I" \y _tip ^,^"^,^;. :,,z ^•-• , 1 ° f ` I ,• +^e^,,.s. ^';. C/'. ^w / ^,•. . .,^ i•; t ^ ^ 15I"^f^ t i r,^^ S ^ t .,ly ^c y^ 1 ^ t ^^ ^ :J j7S^ v (h

tvk

tv.`l --^. "_ . — WO-30 r♦s ^o LEGEND

Crop areas- Cand grassland. and Use Mau of "Ribeirao Preto" Fig.II1 .2 -L ^r region, Sao Paulo State, Brazil •1 "Campo cerrado". k` based on LANDSAT inane interpre-

tation(channe ls S and 7). Natural forest.

o- Gallery forest.

Ig - Artificial forest. ORIGINAE pAGE POOR QUALITY j - Rcads.

ED,-Urban areas.

- State boundary. _^ °<^ ^.:mot, - e ^ ^ .1 - `•

of 100^ L 4

SYPgO >1 - w•- E i .'^. A4 t • n^iY^1.n... ;^i ar^. (_ • -_ ^^; ^ 1 L. / t4,• . ^ rv« 4• ^,t. •`•'^T' _ y'^`^ \ ', . t^ 1..+ `•^,ay.M.inC``^0. 4, `^ i'• ^ 7 L 'Jr^_r ,f,.;V _r" -!" ^ ^ tii..! Ca` tat ^ - S - t N.Y^^f. -••W '`^^'_ j^^ l• ^t^. ' L_a •^N ^tiry ^^^t^ ` ^•_ W e'I `Ta

jy. fYIV,^V \ Y^l. ' t• ` .0 . -^ 11. • r lr • i ^z J .. R i^f{ _ r tra x. !` } \.1 I •.. t • t4N Y_3t^à i r~ î î f r .f i ( r __ ^î ^•^ ^ {tom• .lvq .. i ' " . t ì "^^• { f ^ . rot\^., 4!t• j V 11 4 ♦ , t ),e•.^'."...^„7 ^ ` ^. •rt.' ^"^`.. f r.f \ r ^ y 1.4^" ► ^\,) {+ t" i t^ ^^ ^ J rrl. `•^; } ,4 J l t. )y\.o . M.^'... ? E S`L ♦ j Z A^ C _ I r C+^ .P ^:. ^'^ Yet. t.. N«^• ^ • .'r.. ^( ^.i *^♦ \.',-,^ ----•'_^ F' .. ^ \, _^ } n ,.. lS 3. . ." • 1 ;^:'_ t t 1, \:. :.5' 1 °^( ^ '•' '' i I>,'^ , , W ^ ^ t ^ !I - ' ^^♦ . 1 '^•J ^ ^ F . lyî•^ 5 (^1' if U^ r`^; ^ t.. ^s ♦( (::,..^ ^ f' ll ^ - ., tt ^ . ' wS x r a1 !1_I rat ' f--., ^ lr .. ^ ^ _ . .

^--'.-.:r \ --. u. ^. ^ , 1. -q . - 1 ,; , _"` ^.'--•.^ -.1r _ a" ` /-"Z `1 ^ . i t a.. e.. * ^^/ 11r •v. -.*^ y lv. - _ J - .... Kam.. ^.. , • ^ _ y r •-> / ()) •a' J ^^ Y.. /l,. L•(C ^ _ '_ `(`fit .',F v ^il ,a•^•xit! \ - W ^-Sf p `- C«. tM ! ^r^•

^ alt - ^-t - 151`` ^^_• %^ - -- 111 -^•^ i `^ .Ltr ' t•.. N. 6 a,1i^-' ••r-•tY ~ \It .^ -^. tr. _. .Y i 1 ^ `^^j ^ ,,.i i S12'^0 ^ ^ .yr--^_- • 1 1.{<,^1 ^^ •r..r.^ rµj^ • ttt^ v \44. ) >«. ,., •..a , 'w+R_ `^.^/_.- 3 s — ^ ^^^/' F'4` ^ Vim. ? r: ' " ,.,^- t.. r ., • r -^ _' ` .,,, ^' n

ra r•)o r•s•)o ru'oo

Fig.1II•3 y Soil Map of "Ribeirao Preto" region, Sao Paulo State, based on the Brazil's Commission Soil Ma p (1960). ,,h!GINAL PAGE IS t}F PUM., QUALM LEGEND }} t PV - Ortho Red-Yellow Podzolic Soils. ,H1 - Hydromorphic Soils. PVa - Red Yeilbw Pedlelio toils Piracicaba A - Alluvial Soils. I Variation. Li.b - Lithosol Basalt substratum Phase. POS"; k4d Yellvii Podtbiic Soils Laras Varia Li.gr- Lithosol Granite-Gneiss Substratum ti ori. - Phase. L`' Plii = Pctlzblied Soils on Calcareous Sand - Li.ac'- Lithosol Calcareous Sandstone Subs-

i^VW(r Li'Hb ^^tRiB^lon. tratum Phase. Pml - Podzolized Soils on Calcareous Sand - R - Regosol. stone Narilia Variation. RPV/RLV - REgosol Intergrade to Red-Yellow LR - Terra Rossa. Podzolic Soils and Intergrade to LE - Ortho Dark Red Latosol. Red-Yellow Latosol Undifferentia

ted Soil Group. - -• LEa - Dark Red Latosol Sandy Phase. . LVa - Red Yellow Latosol Sandy Phase.

a. Forest and "Cerr.ad`ao"; b. "Cerrado"; C. "Campo Cerrado".

5 Natural vegetation was considered a poor parameter to make the correlation between soil and vegetation because or man made modifications. This was observed also by Amaral and Aud ; (1972).

3.4 - RELIEF

Based on textural features of the images of channels 6 and 7, tha relief was classified as:

a. slightly ondulating to flat land; b. ondulating land; c. and mountainous.

The results of the interpretation are shown in Figure III.5.

3.5 - DRAINAGE

Figure III.6 shows the drainage pattern of the area. It was obtained from conventional interpretation of channels 5 and 7. Main rivers were well characterized in channel 7. S:cond,-r.y rivers were better displayed in channel 5 because of gallery forest and the vegetation of poorly drained soils.

Table III.1 summarizes quantitatively and qualitatively the data. The analysis of the drainage pattern was separated in homogeneous areas. Table III.2 presents data of correlation and regression analysis of quantitative features disclosing highly significant correlation.

The results obtained w:-. re compared with aerial photograitic results of Franga (1968), Fadel (1972), Masques Filh0 (1972) and Souza (1975). The Computed image indices were similar to those calculated ii, the above investigations when the drainage was not dense. ~ .

j 1'" j j:. 'j ','. ..~ \ J 4

LEGEllD ~ _ Slightly ondul.ted to Fl.t L.nd. Fig. . _ Map of Relief Cl~5ses of -Ribci S Roning to Ondulated Land. III rao Preto" region. Sao Paulo St! o - teo Brazil based on LANDSAT inter pret~tion images (channels 6 and 7 ).

Vim( f ^ , y 1 , •:^: ti^, 1 , . ..`\ f „ .^^ .ti. 1. _ f. 1 r^-. , ^ ;• ^ , •^ ^ ,r

• ^'J `-t'^ } ;;•:...

fit

Y 4 - ` • j"4. -^ ,„^a'..' i \ 1 i 'i,,^fr.' , ^ =. '^ •^ {j ^I + }^ lr. ,• lea

^^.•^J :' u-•,: r ''^ .l • ;.y •lrr,J+'1^^ ^ `^ j r', - ' ^ '. ^ ^' ^ r_ -.^ ^• 1t?,;,S.l•.a J y 1 1 ^, L^, :ti I J JJJ''' l . 1 1^ _ 3i SSA IL

e ••e • ao ^ ' LE ;E!!G ^ ^

Drainage. Ir Fi^.IlI.L - Drainage MaP Of "Ribeirao Preto" r region, Sao Paulo State, bzsed Alluvial Plains. on LANDSAT interpre-tation iira,es

(, cannel s 5 end 7). - l - Lakes. r••t L-_ Urban Arc-,a.-

r, ORIGINAL PAGE LS 'Amw OF POOR QUALITY ao -+ N M to M O ► N of CT O O O O O

M, LAJ N C]

C .-- O CV) 11) t0 .- eT 00 00 O E

d Z O O r Ln C14 F Z Q-. O ^ O M Q^► G O O O O 1- W N O C 0 wE GEam! cv) lLn %0 w co LL O O O O .-

z C3 O cx zo ¢-. c^ ° o o 1--Q^ ►W o 0 0 0 0 N p

CA LO i ^ E C1rtf co co ko C3 E O O O O .-- V- 6-r J U -s ► JJ OC W W 0 W J 0" p C pC Z H co Z M J Q W Q F' Q LLI O 0. p Z L. F-- CC co LLI u ¢ Qi N N Cl-

W J 1 J 1-+ C7 J zCD J 0-4 Zr N J Q J OC t-- cc p aD cj Q z 4 cc 0. p

i, N rJ-r O G N O O N 3 Q Q b MW +. 0.' N O N? N W UJCX UJ C9 O J J p J N d Z E aC JJ— wti+^ "t,^OCC _ 7-NW 0. (nW OQN N wcr - W J WGF QO S4 Q =NQQ S O Zin Q 0. CJ0. OJG. Ll. 0CA Ow ---

t7i C) ce 0. N W

^-^ W O W► N J ca CX 1 mr

-j W C_- i- CC Q: O ~c[ ►-. CV) cf O U J W ^ Ol tNi7 QCU LL. w ti Cl a+ C N ', O M w O a O C '-- 6- N U V r-y N Y w dJ G] rQ d m m w z z z Q d w w aw Cr f U Q r F- U N N_ — N w_ ^ F-- ^I w>- A w ova.. v c wF r w ww^ w omt, c =I co c C crci V) a ww z u ¢ Q Ckf 0: 2 d w d CL

amC'` r N [i_

dd0 x w { - U, W to t F 1-- i w- Y+

t I - 17 -

3.6 - SPECTRAL MEASUREMENTS

Multispectral characteristics of five soil groups were measured in the field (Table III.3) and a statistical analysis was perforuied on the dat4 obtained (Table IIIA). Laboratory analysis was performed for four different soil groups, and Figure III.7 shows the corresponding spectral curves.

The results observed agree with the studies-done by Crown and Pawluk (1974). The spectral responses of the soils under natural and laboratory conditions are a direct function of their physical and chemical properties.

Only three soil groups could be distinguished by the analysis of the spectral response according to the gray levels of the positive,transparencies in the scale of 1:1,000,000 (Fig. III.8). The resu]ts obtained agree with those of Hoffer and Anuta (1972). - 18 -

i 6 J 3 Z N b• C1 p1 Cb O w w w •t ^ Q w w Q tt r C1 Q IL f C r t US r r N N m F OC F- O Cz 4J to N C M a J A L W QL Z t+r1 1^ co to {\ L S w w w w w dJ C r ^D M r N N N !"^ U Y Z F C 3 A C N LL N Y LLI O V H d 12 J 4 F- W V a Z O n N r M V W 2 w • w w w V d r tp 00 M O O H C a S r- r r N t N W U 'C3 a O d C sZ a Z 3 O v+ CD L4 r d ^- 4J r r H C ^ Y v J >- O d tt 1 N 4 v J a 7E t+ z v cn ai co uo V ^ w w w w w O A V 9- S r r N 41 O C C U a d V of L L d V L Y^ y 4 c C C cc to Y ^ G J W O ,r IL J Ln Z O .^ O O 1 v: f- ICF a 0 0 C N F- _ CD w J N 3 1/1 w ^ J u Q C, acc C73 X —j n c a pC W J O 7 l+J Y >- Gti N sr s 1 Q \ O 7 QQ W Q CL W c cn eY N a of

c s TABLE II I .4 ,

STATISTICAL Af.ALYSIS OF THE MEAN REFLECTANCE OF THE FIVE DIFFERENT SOIL GROUPS

j TUKEY'S TEST 1 I F CV ~2 DUNCAN'S TEST (l = 0,05 a = 0,05

I CHANNEL I LR LEa LVa RPV/RLV R *** 16% 0.44 LR LEa LVa RPV/RLV R ! 7 ------I CHANNEL LR LEa LVa RPV/RLV R 17% 0 50 lR LEa LVa RPV/RLV R - *** 1 • - 6 I . CHANNEL LR LEa.. LVa RPV/RLV R *** 15~ LR LEa LVa RPV/RLV R 5 1- 1°·39 --- -

CHANNELl LR LEa LVa RPV/RLV .!! *** 15~ 10.24 LR LEa LVa RPV/RlV R I ------I 4 ,--- 1" -~

1 - Coefficient of variation 2 - Standard deviation -- — . 1 — — — — — ---elo

- 1

.3e AO 42 +4 46 N so 52 .s• .56 56 60 62 94 .66 as .70 72

rlavelength (MICRONS)

SOIL GROUP

Terra Rossa

Red-Yellow Latosol

Dark Red Latosol Sandy Phase

Podzolized Soils on Calcareous Sandstone

Regosol Intergrade to Red-Yellow Podzolic

Soils and Intergrade to Red-Yelloy Latossol

Undifferentiated Soil Group

Regosol

Fig.III,7- Spectral curves of six (6) different soil groups.

IT / ^` ~ 2| . .| . i ' @ ' ^ \ & ' |

^ ^n !

^ ^

` LE a `

LLJ

. ^ e k ^m K ^

x ` | LEa N^U

u zu ao 40 R[LA[lVE VALUES OF TONALITY

REQ}SOL (R)

NMI] TERRA — R05SK /LR\

DARK RED L8TOSOL SANDY PHASE (LEa)

Fig,yll'8 - Graphic presentation of the relative tonality of MSS positive ; transparencies for three soil groups. \

, PAGE IS OF POOR QIJALnTyl

\' ^ ' ! ^ ^ ^ /

^ CHAPTER IV ( _ 4 CONCLUSIONS

To characterize soils, based on LANDSAT images,it has ( been concluded that:

It is necessary to survey annual water balance to know the apparent superficial modifications ir, the images due to soil water content.

Land use, drainaye pattern, relief and tonality measurements are useful^elenents to characterize major soil groups.

Channels 6 and 7 are considered the best to study the relative tonality of different spectral responses of soils. Channels 5 and 7 are considered the best for studying natural vegetation, drainage pattern and land use. Channel 6 is considered the best to study the relief. i

A better contrast of the interpretation elements was { t' obtained in the scale 1:1,000,000 but the scale 1:500,000 was considered more efficient to delineate the interpretative elements.

4 Natural vegetation and land use are factors that should be used carefully when characterizing major soil groups. i

Frequency ratio is the recomended index for use when ± analysing drainage pattern quantitatively !The type of drainage pattern is the most important factor in the qualitative analysis.

It is possible to correlate relief and soil typesin the images.

"Terra Rossa", Dark Red Latosol Sandy phase, Reglsol "were characterized by relative tonality over positive transparencies.