1704 Moaci r Tuzzi n de Moraes et al .

R. Br as. Ci . Sol o, 36:1704-1713

CORRECTION OF RESISTANCE TO PENETRATION BY
PEDOFUNCTIONS AND A REFERENCE SOIL WATER
CONTENT
(1)
Moacir Tuzzin de Moraes
(2)
, Henrique Debiasi
(3)
, J ulio Cezar Franchini
(3)
& Vanderlei
Rodrigues da Silva
(4)
SUMMARY
The soil penetration resistance is an important indicator of soil compaction
and is strongly influenced by soil water content. The objective of this study was to
develop mathematical models to normalize soil penetration resistance (SPR), using
a reference value of gravimetric soil water content (U). For this purpose, SPR was
determined with an impact penetrometer, in an experiment on a Dystroferric Red
Latossol (Rhodic Eutrudox), at six levels of soil compaction, induced by mechanical
chiseling and additional compaction by the traffic of a harvester (four, eight, 10,
and 20 passes); in addition to a control treatment under no-tillage, without chiseling
or additional compaction. To broaden the range of U values, SPR was evaluated in
different periods. Undisturbed soil cores were sampled to quantify the soil bulk
density (BD). Pedotransfer functions were generated correlating the values of U
and BD to the SPR values. By these functions, the SPR was adequately corrected
for all U and BD data ranges. The method requires only SPR and U as input variables
in the models. However, different pedofunctions are needed according to the soil
layer evaluated. After adjusting the pedotransfer functions, the differences in the
soil compaction levels among the treatments, previously masked by variations of
U, became detectable.
I ndex terms: pedotransfer functions, Oxisol, impact penetrometer, soil
compaction.
(1)
Recei ved for publ i cati on i n February 15, 2012 and approved i n September 27, 2012.
(2)
Master Student, Pos-Graduati on Program of Soi l Sci ence, Uni versi dade Federal de Santa Mari a - UFSM. Ci dade Uni versi tri a.
Av. Rorai ma, 1000, Camobi . CEP 97105-300 Santa Mari a (RS). E-mai l : moaci r.tuzzi n@gmai l .com
(3)
Researcher, Embrapa Soja. Rod. Carl os Joo Strass, Di stri to de Warta. Cai xa Postal 231. CEP 86001-970 Londri na (PR). E-
mai l : henr i que.debi asi @embr apa.br ; jul i o.fr anchi n@embr apa.br
(4)
Pr ofessor , Depar tament of Envi r onment and Agr onomy Sci ence, UFSM. Campus de Fr eder i co Westphal en. Li nha Sete de
Setembro, s/n. Rod. BR 386, km 40. CEP 98400-000 Frederi co Westphal en (RS). E-mai l : vanderl ei @ufsm.br
CORRECTI ON OF RESI STANCE TO PENETRATI ON BY PEDOFUNCTI ONS AND A REFERENCE... 1705
R. Br as. Ci . Sol o, 36:1704-1713
RESUMO: CORREO DA RESI STNCI A PENETRAO USANDO UMA
UMI DADE DO SOLO DE REFERNCI A E PEDOFUNES
A resistncia do solo penetrao (SPR) um dos principais indicadores do estado de
compactao do solo; contudo, a SPR altamente influenciada pelo contedo de gua no solo.
O objetivo deste trabalho foi desenvolver modelos matemticos para a correo da SPR para
um valor de umidade gravimtrica (U) de referncia. Para isso, a SPR foi determinada, por
meio de um penetrmetro de impacto, em um experimento instalado sobre um Latossolo
Vermelho distrofrrico, usando seis nveis de compactao do solo, obtidos por meio da
escarificao mecnica e da compactao adicional, pelo trfego de uma colhedora de gros
autopropelida (quatro, oito, 10 e 20 passadas), alm de uma testemunha, a qual foi mantida
sob sistema plantio direto sem escarificao ou compactao adicional. A fim de obter ampla
variao nos valores de U, as avaliaes da SPR foram realizadas em diferentes pocas.
Amostras de solo com estrutura preservada foram coletadas para determinar a densidade do
solo (BD). Foram geradas funes de pedotransferncia relacionando os valores de SPR, de U
e de BD. Usando essas funes, a correo da SPR foi satisfatria para todas as amplitudes de
U e BD. O mtodo requer apenas SPR e U como variveis de entrada dos modelos. No entanto,
so necessrias diferentes equaes, em funo da camada de solo avaliada. A aplicao das
funes de pedotransferncia, obtidas neste trabalho, permite observar diferenas no estado de
compactao do solo entre os tratamentos, que antes no eram detectadas, em funo de
variaes na U.
Termos de indexao: funo de pedotransferncia, Latossolo Vermelho, penetrmetro de
impacto, compactao do solo.
INTRODUCTION
The soi l compacti on l evel can be assessed based on
soil penetration resistance (SPR), which is determined
by penetrometry. Thi s method has some advantages,
e.g., the easi ness and speed i n obtai ni ng data;
i denti fi cati on of compacted l ayers at di fferent depths;
and high correlation with plant root growth (Bengough
et al ., 2011).
Despi te the advantages of usi ng penetrometers,
SPR vari es di rectl y i n functi on of soi l bul k densi ty
(BD) and i nversel y i n functi on of soi l water content
(U) (Busscher, 1990; Bengough et al ., 2001; Vaz et
al ., 2011). Thi s fact l i mi ts compari sons of soi l s of the
same type wi th di fferent water contents, once a smal l
reducti on or i ncrease i n U resul ts i n a l arge i ncrement
or reducti on of SPR (Vaz et al ., 2011), l eadi ng to an
under- or overesti mati on of the soi l compacti on l evel .
Correl ati on between SPR and soi l water content have
al ready been exhausti vel y studi ed; neverthel ess SPR
depends on factor s such as: soi l management
(Busscher et al ., 1997); compacti on l evel (Torres &
Sarai va, 1999; Bengough et al ., 2001; To & Kay, 2005;
Al mei da et al ., 2008); organi c matter content (To &
Kay, 2005); and soi l texture (Al mei da et al ., 2008;
Vaz et al ., 2011).
The most commonl y used functi ons to express
correl ati ons between SPR and gravi metri c soi l water
content (U) were the negati ve exponenti al (Al mei da
et al ., 2008; Si l vei ra et al ., 2010) or negati ve power
functi on (Busscher et al ., 1997; Al mei da et al ., 2008;
Si l vei ra et al ., 2010; Vaz et al ., 2011) equati ons. Vaz
et al . (2002) i ndi cated that the i deal procedure woul d
be to measure U together wi th SPR, and appl y some
type of correction or normalization for a reference value
of soi l water content afterwards. Thi s procedure may
reduce i nterpretati on probl ems of resul ts obtai ned
under di fferent fi el d condi ti ons and soi l management
systems (Busscher et al ., 1997).
One of the earl i est attempts to correct the val ues of
SPR to a reference val ue of soi l water content was
performed by Busscher (1990), who adjusted regression
equati ons usi ng SPR, BD and U data. Busscher et al .
(1997) developed a method to correct SPR as a function
of U, based on the fi rst term of a Tayl or seri es. These
author s, however , adjusted r egr essi on equati ons
cor r el ati ng SPR excl usi vel y wi th U, wi thout
considering other variables, such as the soil compaction
l evel . Thus, normal i zati on of SPR to a reference val ue
of U, through that method, implies in the use of different
r egr es s i on equati ons , i n functi on of the s oi l
management system. According to Almeida et al. (2008),
the correcti on of SPR data needs several equati ons,
based on di fferent condi ti ons of soi l texture and BD.
A correcti on method, based on vol umetri c soi l
water content (), associ ated to the matri c potenti al
of -10 kPa, together wi th the procedure proposed by
Busscher et al . (1997), was proposed by Vaz et al .
(2011). These authors descri bed the need to measure
the BD together with the SPR to determine the values
as a great di sadvantage of the method. I n thi s sense,
Vaz et al . (2011) concl uded that U-based model s coul d
be a l ess compl ex and l abori ous procedure of SPR
correcti on.
The objecti ve of thi s study was to devel op a method
from the procedure proposed by Busscher et al . (1997)
1706 Moaci r Tuzzi n de Moraes et al .
R. Br as. Ci . Sol o, 36:1704-1713
for a Rhodi c Eutrudox, to correct SPR val ues to a
reference val ue of U through pedotransfer equati ons
that correl ate SPR to BD and U.
MATERIAL AND METHODS
Thi s study was carri ed out at the Experi mental
Stati on of Embrapa Soybean, i n Londri na, State of
Paran, Southern Brazi l (l at. 23
o
11 S, l ong. 51
o
11
W; 620 m asl ). The experi ment was carri ed out on a
Dystroferric Red Latossol (Brazilian soil classification)
(Santos et al ., 2006), or Rhodi c Eutrudox (Ameri can
soi l cl assi fi cati on) (Soi l Survey Staff, 2010), under no-
ti l l age (NT) si nce 1996, wi th 731 g kg
-1
cl ay, 146 g kg
-1
si l t, 123 g kg
-1
sand, 18.50 g kg
-1
organi c carbon, and
parti cl e densi ty of 2.96 Mg m
-3
i n the 0-20 cm l ayer.
The aver age sl ope of the exper i mental ar ea i s
0.03 m m
-1
.
To obtai n the SPR cor r ecti on equati ons as
functi on of U, the vari ati on range of the parameters
used for thei r adjustment (SPR, BD, and U) must
be wi de enough. For thi s pur pose, an exper i ment
was establ i shed i n r ows i n August 2010, usi ng a
compl etel y r a n domi zed des i gn , wi th two
repl i cati ons. The treatments were arranged i n pl ots
(2.5 x 20 m), consi sti ng of si x soi l compacti on l evel s:
NT wi th r ecent chi sel i ng (NTCh); NT wi thout
chi sel i ng and no addi ti onal compacti on (NT); and
NT wi th addi ti onal compacti on by harvester traffi c
at four i ntensi ti es, i nduced by four (NTH4), ei ght
(NTH8), 10 (NTH10), and 20 (NTH20) passes over
the same tr ack. The har vester used had a mass of
10.28 Mg and a ti r e-soi l contact pr essur e of 0,23
MPa i n the fr ont axl e. The soi l chi sel i ng was
performed by usi ng a 5-shank chi sel pl ow, reachi ng
a depth of 0.3 m.
Pr i or to the ear l i est SPR assessment and after
tr eatment appl i cati on, the whol e exper i ment was
i r r i gated (i r r i gati on l evel 100 mm), i n or der to
uni for mi ze and r ai se U val ues to over 32.3 %,
cor r espondi ng to the soi l fi el d capaci ty of the
experi mental area, determi ned at a matri c potenti al
of 0.01 MPa i n a pr es s u r e pl ate appar atu s
(Embr apa, 1997). The SPR was measur ed i n ni ne
eval uati ons (two, thr ee, four , seven, ni ne, 11, 14,
23, and 31days after i rri gati on) whi ch, together wi th
the l ack of r ai nfal l i n the exper i mental per i od,
wi dened the var i ati on r ange of U val ues. The SPR
was determi ned i n the soi l l ayers 5.5-10.5 and 13.5-
18.5 cm, wi th an i mpact penetr ometer (model
I AAPl an al s u car -Stol f) (Stol f, 1991), u s i n g a
130 mm
2
base ar ea and 30
o
ci r cul ar stai nl ess steel
con e (As abe, 2010). Th e SPR r eadi n gs wer e
per for med at ei ght poi nts, 15 cm away fr om each
other , al ong a l i ne (tr ansect) tr ansver sal to the
har vester tr ack and chi sel pl ow passes. I n each
assessment, two r epl i cati ons (tr ansects) wer e used
per compacti on l evel . Next to each tr ansect, two
soi l sampl es were col l ected (l ayers 5.5-10.5 and 13.5-
18.5 cm), to deter mi ne U, accor di ng to Embr apa
(1997).
Undi stur bed soi l sampl es wer e col l ected i n
stai nl ess steel cores (hei ght 5 cm x i nternal di ameter
5 cm), hori zontal l y i nserted by means of a hydraul i c
jack, i n the center of the l ayers 5.5-10.5 and 13.5-
18.5 cm, on the wal l of trenches opened i n each pl ot.
For each treatment and soi l l ayer, 24 soi l cores were
sampled, totaling 288 samples. I n the laboratory, these
cores were anal yzed for BD, and U equi val ent to the
soi l fi el d capaci ty (0.01 MPa) by means of a pressure
pl ate apparatus (Embrapa, 1997). The permanent
wi l ti ng poi nt was deter mi ned i n di stur bed soi l
sampl es, accordi ng to the method descri bed by Kl ei n
et al . (2006), usi ng a thermocoupl e psychrometer
(Decagon, model WP4-T).
The SPR val ues were adjusted to U val ues by an
equati on of the potenti al type used by Busscher et al .
(1997). To faci l i tate l ater computati ons, the equati on
was l i neari zed (Equati on 1):
RP = b + a Ln(U) (1)
where b and a, are empi ri cal l y adjusted parameters
of the model s; and Ln(U) = natural l ogari thm of
gravi metri c soi l water content.
Functi ons expressi ng the correl ati on between SPR
and U were esti mated for each treatment (compacti on
l evel s) and soi l l ayer. Thi s means that, correspondi ng
to each mean BD of the di fferent treatments and
l ayers, there i s a val ue for parameter a, whi ch i s the
angul ar coeffi ci ent of equati on 1 (i .e., the fi r st
derivative) and therefore represents the variation rate
of SPR wi th LnU. I n thi s sense, the extent of SPR
vari ati on i n functi on of U i ncreases wi th i ncreasi ng
BD (Bengough et al ., 2001; To & Kay, 2005; Al mei da
et al ., 2008). Therefore, the next step of the method,
whi ch represents a major advance i n rel ati on to the
correcti on procedure proposed by Busscher et al .
(1997), was to r el ate the modul e of the angul ar
coeffi ci ent of the functi ons that r epr esent the
rel ati onshi p between SPR and LnU (| a| ) wi th the
mean BD measur ed per tr eatment, accor di ng to
equati on (2):
| a| = c BD
d
(2)
where c and d are empi ri cal l y adjusted parameters of
the model s.
The SPR was corrected by the fi rst term of a Tayl or
series (Equation 3) (Busscher et al., 1997). The U value
of 27 % was used as r efer ence to cor r ect SPR,
correspondi ng to the center of the fri abi l i ty range, as
determi ned by Torres & Sarai va (1999) for the same
soi l type, wi th a soi l management si mi l ar to that used
i n the experi ment of thi s study.
SPR
corr.
= SPR
read
- [| a| (LnU
ref. -
LnU
read
)] (3)
CORRECTI ON OF RESI STANCE TO PENETRATI ON BY PEDOFUNCTI ONS AND A REFERENCE... 1707
R. Br as. Ci . Sol o, 36:1704-1713
wher e SPR
cor r .
= soi l r esi stance to penetr ati on
corrected i n functi on of the reference val ue of the
gravi metri c soi l water content (U
ref.
); SPR
read
= soi l
r es i s tan ce to pen etr ati on r ead i n th e fi el d;
Ln U
r ef.
= n atu r al l ogar i th m of U
r ef.
(27 %);
LnU
read
= natural l ogari thm of U, of the same si te
and soi l l ayer as the SPR assessment.
So far, the method proposed to correct the SPR i n
functi on of U, requi red BD data to determi ne | a| .
However , when SPR i s used as i ndi cator of soi l
compaction in the field, the assessment of BD is usually
i nfeasi bl e, once the quanti fi cati on method i s l abori ous
and ti me-consumi ng. To el i mi nate the need of
determining BD, pedotransfer functions were adjusted
to estimate this property in function of SPR
read
, specific
for si x i nterval s of U val ues i n the 5.5-10.5 cm l ayer
and four i nterval s i n 13.5-18.5 cm (Tabl e 1). The U
i nterval s were determi ned so as to meet the fol l owi ng
cri teri a, i n both soi l l ayers: i ) a mi ni mum number of
fi ve poi nts per soi l water content i nterval ; i i ) i n each
interval, all treatments should be present for a sufficient
vari ati on range of BD and SPR; and i i i ) determi nati on
coeffi ci ent (r
2
) of pedofuncti ons > 0.70.
To val i date the proposed method, the pedotransfer
functi ons were appl i ed to SPR data measured i n
another experi ment by Torres & Sarai va (1999), i n
the l ayers 5.5-10.5 and 13.5-18.5 cm of a very cl ayey
Rhodi c Eutr udox under no-ti l l age at four soi l
compacti on l evel s eval uated over ti me, total i ng 13 U
val ues. I t i s worth rememberi ng that thi s data set
was not used for the pedofuncti ons adjustment, but
onl y to val i date the method. For thi s purpose, the SPR
data at U of 27 %, deter mi ned at di ffer ent soi l
compacti on l evel s usi ng equati ons rel ati ng SPR to U
adjusted wi th the data obtai ned by Torres & Sarai va
(1999), were compared to SPR data corrected for the
same U value, by the method proposed here, by means
of l i near regressi on anal ysi s.
Resul ts were subjected to ANOVA and treatment
means compared by the Tukey test, at 5 % probability,
usi ng the Stati sti cal Anal ysi s System (SAS, 2002).
The same software was used for regressi on anal ysi s
to adjust the pedotransfer functi ons.
RESULTS AND DISCUSSION
I n both soi l l ayers, BD was si gni fi cantl y hi gher i n
the treatments wi th addi ti onal soi l compacti on after
harvester traffi c than i n the NT (Tabl e 2). I n contrast,
soi l ti l l i ng by chi sel i ng resul ted i n l ower BD than i n
the other treatments. I n the 5.5-10.5 cm l ayer, BD
var i ed fr om 1.10 Mg m
-3
(NTCh) to 1.50 Mg m
-3
(NTH20), whi ch i s near the maxi mum BD esti mated
by th e Nor mal Pr octor Tes t for th i s Ox i s ol ,
corresponding to 1.53 Mg m
-3
(Torres & Saraiva, 1999).
Si mi l arl y, i n the l ayer 13.5-18.5 cm, BD vari ed from
1.10 Mg m
-3
(NTCh) to 1.42 Mg m
-3
(NTH20).
Therefore, the BD vari ati on range was l arge i n both
soi l l ayers, whi ch i s a basi c requi rement for adequate
pedofuncti on adj ustments to cor r ect SPR to a
reference U val ue.
I n the 5.5-10.5 cm l ayer of treatment NTH20, the
introduction of the penetrometer into the soil led to the
formati on of verti cal cracks bel ow the cone, resul ti ng
in low SPR values (data not presented), incoherent with
the hi gh BD measured i n thi s l ayer and treatment
(Tabl e 2). At hi gh BD and l ow U val ues, To & Kay
(2005) also observed that the penetrometer movement
i n the soi l forms smal l verti cal craks ri ght bel ow the
cone, reducing the SPR values. Therefore, the SPR data
obtai ned i n NTH20, l ayer 5.5-10.5 cm, wer e not
consi dered for the establ i shment of the pedotransfer
functi ons. Thi s probl em was not observed i n the 13.5-
18.5 cm l ayer, possi bl y as a consequence of l ower BD
than i n the 5.5-10.5 cm l ayer. Thus, resul ts obtai ned
i n treatment NTH20, l ayer 13.5-18.5 cm, were used
i n the adjustment of pedotransfer functi ons.
The assessments of SPR i n di ffer ent per i ods
resul ted i n a wi de vari ati on range of U, whi ch i s al so
a requi rement for the adjustment of pedotransfer
functi ons. Val ues of U vari ed between 34.1 and 21.1
% (5.5-10.5 cm), and between 34.8 and 23.6 % (13.5-
18.5 cm). Wi thi n these vari ati on ranges, mi ni mum
and maxi mum val ues were si mi l ar to the U val ues
associated with the permanent wilting point (1.5 MPa)
and the fi el d capaci ty (0.01 MPa), correspondi ng, i n
the mean among treatments, to 24.8 and 32.3 %,
respecti vel y. The l arge vari ati on ampl i tude for BD
and U (Table 2) resulted in a wide range of SPR values
(0.57 - 21.66 MPa i n the 5.5-10.5 cm, and 0.74 -
14.31 MPa i n the 13.5-18.5 cm).
Interval of U (%) Equation
(1)
No. of equation
5.5 - 10.5 cm
I f, 21.1 < U 23.1 BD = e RP
f
4
I f, 23,1 < U 25.1 BD = g + h RP 5
I f, 25.1 < U 27.1 BD = i + j RP 6
I f, 27.1 < U 29.1 BD = k RP
l
7
I f, 29.1 < U 31.5 BD = m RP
n
8
I f, 31.5 < U 34.5 BD = o RP
p
9
13.5 - 18.5 cm
I f, 23.1 < U 26.3 BD = q RP
r
10
I f, 26.3 < U 29.1 BD = s RP
t
11
I f, 29.1 < U 32.1 BD = u RP
v
12
I f, 32.1 < U 35.1 BD = x RP
z
13
Table 1. Equations to estimate the soil bulk density
(BD) using the soil resistance to penetration
(SPR) as independent variable, for different soil
layers and intervals of gravimetric soil water
content (U), on a Rhodic Eutrudox
(1)
e, f, g, h, i, j, k, l, m, n, o, p, q, r, s, t, u, v, x, and z are empi ri c
parameters to adjust the model s.
1708 Moaci r Tuzzi n de Moraes et al .
R. Br as. Ci . Sol o, 36:1704-1713
The correlation between SPR and U was expressed
sati sfactori l y by equati on (1), si nce the adjustment
was stati sti cal l y si gni fi cant (p<0.01) and r
2
val ues
were > 0.90, for al l treatments and soi l l ayers (Fi gure
1). When U i ncreased, the SPR di fferences among
treatments decreased (Fi gure 1), i n agreement wi th
resul ts of Torres & Sarai va (1999). The di fferences
among treatments for SPR practi cal l y di sappeared
when U approached fi el d capaci ty (32.3 %; LnU 3.47)
(Fi gure 1). Thus, determi ni ng SPR at fi el d capaci ty,
as recommended by the standard ASABE EP542
(ASABE, 2006), i s i nappropri ate when the objecti ve i s
to use the SPR as an i ndi cator of the effects of traffi c
and soi l management on the compacti on l evel .
The parameter | a| of the regressi on equati ons
rel ati ng SPR to LnU i ncreased wi th the i ncreasi ng
compacti on l evel i n each treatment (Fi gure 1), i .e., the
higher the compaction level, the higher the increase in
SPR with decrease in U, as already described by Torres
& Sarai va (1999) and Bengough et al . (2001). From
thi s observati on, a power functi on was fi t expressi ng
the var i ati on of | a| wi th the mean BD of each
treatment (Fi gure 2). The use of thi s equati on al l owed
estimating | a| at different BD values, which were then
used i n equati on (3). The val ues of | a| were hi gher i n
the 13.5-18.5 cm than the 5.5-10.5 cm layer for the entire
BD vari ati on range (Fi gure 1), i ndi cati ng that SPR i s
more sensitive to U variation in deeper soil layers. This
resul t al so shows that the correcti on of SPR to a
reference U val ue, through thi s method, requi res
speci fi c pedotransfer functi ons for each soi l l ayer.
The correcti on of SPR by equati on (3) and usi ng
| a| , estimated for each treatment using the BD values
obtai ned i n the fi el d, was effi ci ent for the soi l l ayers
5.5-10.5 cm (Figure 3a,b,c,d,e) and 13.5-18.5 cm (Figure
3f,g,h,i ,j,l ). I n al l cases, the rel ati onshi p between
SPR
read
and U was a negative-exponent power function.
After correcti on, SPR
corr
di d not vary i n functi on of U
(p>0.05). Li kewi se, the rel ati onshi p between SPR
corr
and U gave ri se to a hori zontal strai ght l i ne, paral l el
to the absci ssa axi s, whi ch cr osses the SPR
r ead
approxi matel y at the reference val ue of U (U
ref
),
equi val ent to 27 %. Thus, wi thi n the enti re vari ati on
i nterval of U, the SPR
corr
val ue was si mi l ar to the
SPR
read
val ue, obtai ned i n the fi el d, at 27 % of U.
To el i mi nate the need for determi ni ng BD i n the
fi el d to esti mate | a| , whi ch woul d l i mi t the extensi ve
use of the method pr oposed her e, pedotr ansfer
functi ons wer e adjusted to esti mate BD i n r el ati on
to SPR
r ead
. These functi ons wer e speci fi c for si x U
i nter val s i n the 5.5-10.5 cm l ayer (Fi gur e 4a), and
for four U i nterval s i n the 13.5-18.5 cm l ayer (Fi gure
4b). The model r el ati ng SPR to BD and , pr oposed
by Busscher (1990), was not used, once the BD coul d
not be r el i abl y esti mated at extr eme SPR and U
val ues. The use of U i nstead of most l i kel y reduced
the model pr eci si on i n si tuati ons of hi gh var i ati ons
i n SPR and U.
I n both soi l l ayers and for most U i nterval s, the
r el ati onshi p SPR x BD was better expr essed by
exponenti al -type model s (Fi gure 4), agreei ng wi th
resul ts obtai ned by Busscher (1990). Neverthel ess, i n
the 5.5-10.5 cm l ayer and for U ranges between 23.1
and 25.1 % as wel l as between 25.1 and 27.1 %, the
best fi tti ng was obtai ned by the l i near model (Fi gure
4a). Excepti ng the U i nterval between 32.1 and 35.1 %
Treatment No. Minimum Mean Maximum Median Modal Standard deviation CV (%)
5.5-10.5 cm
NTCh
(1)
24 0.95 1.10 Ea** 1.28 1.11 1.04 0.073 6.67
NT 24 1.16 1.25 Da 1.32 1.25 1.25 0.044 3.50
NTH4 24 1.28 1.36 Ca 1.46 1.36 1.36 0.038 2.84
NTH8 23 1.38 1.43 Ba 1.47 1.44 1.47 0.030 2.11
NTH10 24 1.36 1.45 Ba 1.52 1.46 1.46 0.040 2.86
NTH20 23 1.43 1.50 Aa 1.53 1.50 1.52 0.030 2.01
13.5-18.5 cm
NTCh 24 0.98 1.10 Da 1.23 1.11 1.15 0.069 6.32
NT 24 1.16 1.25 Ca 1.32 1.26 1.27 0.043 3.42
NTH4 24 1.28 1.34 Bb 1.43 1.33 1.33 0.036 2.69
NTH8 24 1.29 1.36 Bb 1.42 1.36 1.35 0.026 1.93
NTH10 24 1.26 1.38 Bb 1.47 1.38 1.44 0.580 4.18
NTH20 24 1.34 1.42 Ab 1.49 1.42 1.43 0.048 3.42
Table 2. Descriptive statistics for the soil bulk density of a Rhodic Eutrudox, assessed at different compaction
levels and in two soil layers
(1)
No-ti l l age wi th soi l chi sel i ng (NTCh), No-ti l l age wi thout chi sel i ng or addi ti onal compacti on (NT), NT wi th addi ti onal compacti on
by 4 (NTH4), 8 (NTH8), 10 (NTH10) and 20 (NTH20) harvester passes. **Treatments fol l owed by the same upper case l etter i n
the same soi l l ayer, or l ower case l etter i n soi l l ayers of the same treatment, do not di ffer from each other stati sti cal l y by the
Tukey test at 5 %.
CORRECTI ON OF RESI STANCE TO PENETRATI ON BY PEDOFUNCTI ONS AND A REFERENCE... 1709
R. Br as. Ci . Sol o, 36:1704-1713
(pl asti c soi l consi stency) i n the 13.5-18.5 cm l ayer
(Fi gure 4b), the r
2
val ues of the other fi tted functi ons
were > 0.93 (Fi gure 4), i ndi cati ng a hi gh preci si on i n
esti mati ng BD, usi ng SPR as i ndependent vari abl e.
I ndependentl y of the soi l l ayer, the SPR
corr
val ues
usi ng the BD esti mated by the pedotransfer functi ons
(Fi gure 4) were si mi l ar to those obtai ned usi ng the
BD observed at each soi l compacti on l evel (Fi gure 3).
The rel ati onshi p between LnU and SPR
corr
usi ng | a| ,
determi ned from the esti mated BD (Fi gure 3), was
al so represented by a strai ght l i ne, paral l el to the
abs ci s s as ax i s , cu tti n g th e SPR
r ead
x L n U
appr ox i matel y at th e v al u es of U
r ef
(27 %);
demonstr ati ng that the method was effi ci ent i n
el i mi nati ng the effect of U on SPR. I n thi s way, the
correcti on of SPR by U can be performed wi thout
determi ni ng BD i n the fi el d.
To val i date the proposed method, the pedotransfer
functi ons wer e appl i ed to SPR data obtai ned at
different soil compaction levels and U values by Torres
& Sarai va (1999), i n previ ous research of the same
soi l type and l ayers used i n thi s study (Fi gure 5). I n
al l si tuati ons, SPR was corrected for the val ues of
U
ref
(27 %), where | a| was determi ned by means of
BD predi cted usi ng SPR
read
as i ndependent vari abl e
i n the pedotransfer functi ons (Fi gure 4).
RPC
RFC
20
15
10
5
0
20
15
10
5
0
0.0 3.0 3.2 3.4 3.6
(a)
(b)
S
o
i
l

p
e
n
e
t
r
a
t
i
o
n

r
e
s
i
s
t
a
n
c
e


-

S
P
R
.

M
P
a
Ln gravimetric soil water content, LnU
X
NTCh NT NTH4 NTH8 NTH10 NTH20 X
X
X
X
X
X
SPR = 9.68 - 2.55 LnU r = 0.94**
SPR = 54.28 - 14.98 LnU
2
r = 0.96**
SPR = 81.94 - 22.54 LnU r = 0.94**
SPR = 95.78 - 26.37 LnU r = 0.94**
SPR = 154.35 - 43.55 LnU r = 0.95**
2
2
2
2
SPR = 23.595 - 6.47 LnU r = 0.98**
SPR = 62.986 - 17.11 LnU
2
r = 0.91**
SPR = 94.261 - 26.05 LnU r = 0.91**
SPR = 118.31 - 32.92 LnU r = 0.99**
SPR = 135.05 - 37.69 LnU r = 0.98**
SPR = 149.28 - 41.75 LnU r = 0.97**
2
2
2
2
2
X
Figure 1. Variation of soil penetration resistance with gravimetric soil water content at each soil compaction
level (NTCh =no-tillage with chiseling; NT =no-tillage without chiseling or additional compaction; NTH4,
NTH8, NTH10 and NTH20 =NT with additional compaction by four, eight, 10, and 20 harvester passes,
respectively), in the layer 5.5-10.5 cm (a) and 13.5-18.5 cm (b), in a Rhodic Eutrudox. **Statistically significant
equations (F test, p<0.01). RPC =range of plastic consistency; RFC =range of friable consistency of soil.
Figure 2. Variation of | a| with the soil bulk density assessed in the field in the 5.5-10.5 and 13.5-18.5 cm layer,
in a Rhodic Eutrudox. **Statistically significant equations (F test, p<0.01).
50
40
30
20
10
0
0.0 1.1 1.2 1.3 1.4 1.5
|
|
a
|a| = 3.09 BD r = 0.99**
13.5 - 18.5 cm
7.55 2
|a| = 1.27 BD r = 0.94**
5.5 - 10.5 cm
9.29 2
Soi l bul k densi ty - BD, Mg m
-3
1710 Moaci r Tuzzi n de Moraes et al .
R. Br as. Ci . Sol o, 36:1704-1713
Pri or to correcti on, the SPR
read
i n the 5.5-10.5 cm
and 13.5-18.5 cm l ayer decreased wi th the i ncrease of
U, fol l owi ng a negati ve-exponent power functi on
(Fi gure 5). However, vari ati on of SPR
corr
wi th U was
not statistically significant in both soil layers, resulting
in a straight line, approximately parallel to the abscissa
axi s. The di spersi on of some SPR
corr
val ues, observed
al ong the strai ght l i ne of the regressi on SPR
corr
x U,
mai nl y i n the 5.5-10.5 cm l ayer (Fi gure 5a,b,c,d), may
be expl ai ned by vari ati ons i n the soi l compacti on l evel
at each point where SPR was measured. The existence
of thi s di spersi on i s i mpor tant, showi ng that the
Figure 3. Correction of soil penetration resistance, in function of gravimetric soil water content in a Rhodic
Eutrudox, in the 5.5-10.5 cm layer: a) NTCh =no-tillage with chiseling; b) NT =no-tillage without chiseling
or additional compaction; c) NTH4 = NT with additional compaction by four harvester passes; d)
NTH8 = NT with additional compaction by eight harvester passes; e) NTH10 = NT with additional
compaction by 10 harvester passes; and in the 13.5-18.5 cm layer: f) NTCh; g) NT; h) NTH4; i) NTH8; j)
NTH10; l) NTH20 = NT with additional compaction by 20 harvester passes. **Significant equations (F
test, p<0.01);
ns
non-significant. SPR
corr
with BD
obs
=SPR corrected in function of U by using the soil bulk
density observed in the field; SPR
corr
with BD
est
= SPR corrected in function of U based on soil bulk
density estimated by the pedotransfer functions shown in figure 4.
2.0
1.6
1.2
0.8
10
8
6
4
2
0
15
12
6
6
3
0
0.0
(a) (b) (c)
(d) (e)
r = 0.89**
2
r = 0.88**
2
r = 0.96**
2
r = 0.93**
2
SPR = 812 U
read
-1.98
S = 138783 U
read
-3.13
PR S = 38655 U
lida
-2.60
PR
r = 0.87**
2
15
12
9
6
3
0
5
4
3
2
1
0
15
12
9
6
3
0
15
12
9
6
3
0
15
12
9
6
3
0
10
8
6
4
2
0
15
12
9
6
3
0
25
20
15
10
5
S = 5900 U
read
-2.69
PR
S = 1651851 U
read
-3.65
PR
0
S
read
PR
S with BD
corr. read
ns
PR
r = 0.98**
2
SPR
read
= 1699189 U
-4.10
SPR
read
= 514986 U
-3.42
r = 0.95**
2
(g)
SPR
read
= 740246 U
-3.48
r = 0.83**
2
(h)
SPR
read
= 9497554 U
-4.21
SPR
read
= 15340387 U
-4.32
SPR
read
= 369543583 U
-5.24
r = 0.97**
2
r = 0.97**
2
r = 0.97**
2
(f)
(i) (j) (l)
0 21 24 27 30 33 36 0 21 24 27 30 33 36 0 21 24 27 30 33 36
Gravimetric soil water content - U, %
S
o
i
l

p
e
n
e
t
r
a
t
i
o
n

r
e
s
i
s
t
a
n
c
e

-

S
P
R
,

M
P
a
S with BD
corr. est
ns
PR
r ead
CORRECTI ON OF RESI STANCE TO PENETRATI ON BY PEDOFUNCTI ONS AND A REFERENCE... 1711
R. Br as. Ci . Sol o, 36:1704-1713
method proposed here mi ni mi zes the effect of U on
SPR, but wi thout changi ng the sensi ti vi ty of that
vari abl e to the soi l compacti on l evel .
The mean SPR
corr
of each soi l compacti on l evel and
l ayer was l i nearl y rel ated wi th the SPR at U of 27 %,
estimated by the equations fitted to the data measured
i n the fi el d by Torres & Sarai va (1999) (Fi gure 6). I n
both soi l l ayer s (5.5-10.5 and 13.5-18.5 cm), the
deter mi nati on coeffi ci ent of the l i near equati ons
representi ng the rel ati onshi p between SPR
corr
and
1.5
1.4
1.3
1.2
1.1
1.5
1.4
1.3
1.2
1.1
0
(a)
(b)
5
Soil penetration resistance - SPR, MPa
10 15 20
S
o
i
l

b
u
l
k

d
e
n
s
i
t
y

-

B
D
,

M
g

m
-
3
0.0
0.0
0.1303
0.1360
0.1479
0.1427
Figure 4. Pedotransfer functions to estimate soil bulk density in function of soil penetration resistance, for
different ranges of gravimetric soil water content in the 5.5-10.5 cm layer (a) and 13.5-18.5 cm layer (b),
in a Rhodic Eutrudox soil. **Statistically significant equations (F test, p<0.01).
15
12
9
6
3
0
15
12
9
6
3
0
0 20 25 30 35 40 0 20 25 30 35 40 0 20 25 30 35 40 0 20 25 30 35 40
Gravimetric soil water content - U, %
S
o
i
l

p
e
n
e
t
r
a
t
i
o
n

r
e
s
i
s
t
a
n
c
e

-

S
P
R
,

M
p
a
(a)
(e)
(f)
(g) (h)
(b)
(c)
(d)
SPR with BD
corr. est.
ns
SPR
read
SPR = 40478 U
read
-2.78
SPR = 30819 U
read
-2.62
SPR = 594038 U
read
-3.50
SPR = 174446 U
read
-3.07
SPR = 5124935 U
read
-4.08
r
2
= 0.73**
r
2
= 0.70**
r
2
= 0.87**
r
2
= 0.84**
SPR = 5877256 U
read
-4.16
SPR = 280890 U
read
-3.27
SPR = 1212356 U
read
-3.77
r
2
= 0.91** r
2
= 0.91** r
2
= 0.98**
r
2
= 0.86**
Figure 5. Variation of soil penetration resistance read in the field (SPR
read
) and corrected (SPR
corr
) with the
gravimetric soil water content, in a Rhodic Eutrudox at different compaction levels, in the 5.5-10.5 cm
layer: a) 1.13 Mg m
-3
; b) 1.30 Mg m
-3
; c) 1.30 Mg m
-3
; d) 1.33 Mg m
-3
; and in the 13.5-18.5 cm layer: e)
1.22 Mg m
-3
; f) 1.28 Mg m
-3
; g) 1.29 Mg m
-3
; h) 1.31 Mg m
-3
. Original data obtained by Torres & Saraiva
(1999). **Statistically significant equations (F test, p<0.01);
ns
non-significant.
1712 Moaci r Tuzzi n de Moraes et al .
R. Br as. Ci . Sol o, 36:1704-1713
8
7
6
5
4
S
P
R
i
n

2
7

%

U
c
o
r
r
.
SPR in 27 % U
read.
0 4 5 6 7 8 0 4 5 6 7 8
(a) (b)
SPR = 0.464 + 0.922 SRP
Corr. red
SPR = 0.429 + 0.966 SRP
Corr. red
r = 0.97**
2
r = 0.97**
2
Strainght-line 1:1
Figure 6. Relationship between the corrected value of soil penetration resistance (SPR
corr
) and soil
penetration resistance observed at the gravimetric soil water content (U) of 27 % (SPR
read
), for the soil
layers 5.5-10.5 cm (a), and 13.5-18.5 cm (b), of a Rhodic Eutrudox. Original data obtained by Torres &
Saraiva (1999). **Statistically significant equations (F test, p<0.01).
SPR observed at 27 % of U was hi gh (0.97), very cl ose
to the 1:1 l i ne. These resul ts proved that the method
of correcti on, exempti ng from fi el d determi nati on of
BD, was effi ci ent i n mi ni mi zi ng the effect of U
vari ati on on the SPR val ues when appl i ed to a data
set whi ch, al though obtai ned for the same soi l type
and l ayer , di ffer s fr om the data used to fi t the
pedotransfer functi ons.
The purpose of correcti ng SPR data i s to di mi ni sh
the effect of U on the i nterpretati on of resul ts. Vaz et
al . (2002) stated that, despi te the SPR readi ngs taken
between fi el d capaci ty and permanent wi l ti ng poi nt,
some type of correcti on i s necessary. Accordi ng to
these authors, the vari ati on of U wi thi n thi s range
causes a l arge vari ati on i n SPR, i n the order of 5 MPa.
Thi s var i ati on i s suffi ci ent to cause si gni fi cant
di storti ons i n di agnoses of the soi l compacti on l evel
based on SPR data.
I n thi s study, the SPR
read
i n the 5.5-10.5 cm l ayer
vari ed from 2 to 8 MPa for the NT (BD=1.25 Mg m
-3
)
(Fi gure 3b), and from 4 to 22 MPa for the NTH10
(BD=1.42 Mg m
-3
) (Fi gure 3e). Thus there was an
overl appi ng between the SPR val ues obtai ned for the
treatments NT and NTH10, i n a way that vari ati ons
i n U may l ead to a mi si nterpretati on of resul ts, as for
exampl e, that SPR i n the NT i s equal or hi gher than
i n the NTH10 treatment. After the correcti on, the
SPR val ues vari ed from 4 to 5.5 MPa i n NT and from
9 to 13 MPa i n NTH10. A si mi l ar behavi or was
observed i n the 13.5-18.5 cm l ayer, where SPR
read
vari ed from 2.5 to 7.5 MPa i n the NT (Fi gure 3g),
and from 3.2 to 11.5 MPa i n the NTH10 treatment
(Figure 3j). The overlapping of SPR values in the 13.5-
18.5 cm was hi gher than i n the 5.5-10.5 cm l ayer,
increasing the probability of erroneous interpretations,
due to eventual vari ati ons of U between treatments.
However, after correcti on, the SPR vari ed between
6.1 to 7.1 MPa i n the treatment NT and from 9.5 to
11.5 MPa i n the treatment NTH10, thus el i mi nati ng
overl appi ng of the vari ati on ranges and mi ni mi zi ng
the ri sk of i nadequate i nterpretati ons. I t i s i mportant
to emphasi ze that the vari ati ons i n SPR
corr
, wi thi n
the same treatment and soi l l ayer, occur mai nl y due
to spati al vari abi l i ty of soi l compacti on l evel s, once
the correcti on was based on the mean BD for each
treatment.
I n rel ati on to other methods of correcti ng SPR by
the soi l water content, the procedure proposed here
has some i mportant advantages. For exampl e, the
pr oposed method can be used at di ffer ent soi l
compacti on l evel s due to the adjustment of regressi on
equati ons to esti mate | a| , usi ng BD as i ndependent
vari abl e, whi ch was not consi dered i n the ori gi nal
method of Busscher et al . (1997). Addi ti onal l y, the use
of pedotransfer functi ons to esti mate BD from the
SPR
read
al l ows a sati sfactory correcti on of SPR to a
reference U val ue wi thout the need of quanti fyi ng BD
i n the fi el d, whi ch woul d hamper the use of thi s
method. Thi s represents an i mportant advantage over
the method of SPR correcti on based on the val ue at
a matri c potenti al of -0,01 MPa, as proposed by Vaz
et al . (2011), whi ch requi res the fi el d esti mati on of
BD. I n thi s context, the use of U i nstead of i s another
characteri sti c of the method proposed here, whi ch
al l ows the correcti on of SPR wi thout the need to
determi ne BD.
Sti l l , i t i s i mpor tan t to con s i der th at th e
pedotr an s fer fu n cti on s , wh i ch r epr es en t th e
rel ati onshi p between SPR and BD i n di fferent U
ranges; SPR and U; and | a| and BD, are speci fi c for
the l ayers and the soi l condi ti ons i nherent to thi s
experi ment. I n other si tuati ons, the fi tti ng coeffi ci ents
or even the model s ar e expected to be di ffer ent.
Therefore, these adjustments are a condi ti on for the
appl i cati on of thi s method to soi l s and l ayers di fferent
from those used i n thi s study.
CONCLUSIONS
1. The proposed method al l ows correcti ng SPR to a
reference val ue of U, wi thout affecti ng the sensi ti vi ty
of the i ndi cator to the soi l compacti on l evel .
CORRECTI ON OF RESI STANCE TO PENETRATI ON BY PEDOFUNCTI ONS AND A REFERENCE... 1713
R. Br as. Ci . Sol o, 36:1704-1713
2. Wi th the correcti on of SPR to a reference val ue
of U, di ffer ences among tr eatments, pr evi ousl y
masked by vari ati ons of U, became detectabl e.
LITERATURE CITED
ALMEI DA, C.X.; CENTURI ON, J.F.; FREDDI , O.S.;
FALQUETO JORGE, R. & BARBOSA, J.C. Funes de
pedotr ansfer nci a par a a cur va de r esi stnci a do sol o
penetrao. R. Bras. Ci . Sol o, 32:2235-2243, 2008.
AMERI CAN SOCI ETY OF AGRI CULTURAL AND
BI OLOGI CAL ENGI NEERS - ASABE. Pr ocedur es for
usi ng and r epor ti ng data obtai ned wi th the soi l cone
penetrometer. ASABE Standard EP542. St. Joseph, 2006.
AMERI CAN SOCI ETY OF AGRI CULTURAL AND
BI OLOGI CAL ENGI NEERS - ASABE. Soi l con e
penetrometer. ASABE Standard S313.3. St Joseph, 2010.
BENGOUGH, A.G.; CAMPBELL, D.J. & OSULLI VAN, M.F.
Penetr ometer techni ques i n r el ati on to soi l compacti on
and root growth. I n: SMI TH, K.A. & MULLI NS, C.E., eds.
Soi l and envi ronmental anal ysi s. 2.ed. New York, Marcel
Dekker, 2001. p.377-403.
BENGOUGH, A.G.; McKENZI E, B.M.; HALLET, P.D. &
VALENTI NE, T.A. Root el ongati on, water str ess, and
mechani cal i mpedance: A revi ew of l i mi ti ng stresses and
benefi ci al root ti p trai ts. J. Exp. Bot., 62:59-68, 2011.
BUSSCHER, W.J. Adjustment of fl at-ti pped penetr ometer
resi stance data to a common water content. Trans. Am.
Soc. Agri c. Eng., 3:519-524, 1990.
BUSSCHER, W.J.; BAUER, P.J.; CAMP, C.R. & SOJKA, R.E.
Correcti on of cone i ndex for soi l water content di fferences
i n a coastal pl ai n soi l . Soi l Ti l l age Res., 43:205-217, 1997.
EMPRESA BRASI LEI RA DE PESQUI SA AGROPECURI A -
EMBRAPA. Manual de mtodos de anl i se de sol o. Ri o
de Janei ro, 1997. 212p.
KLEI N, V.A.; REI CHERT, J.M. & REI NERT, D.J. gua
di spon vel em um Latossol o Vermel ho argi l oso e murcha
fi si ol gi ca de cul turas. R. Bras. Eng. Agr c. Amb., 10:646-
650, 2006.
SAS LEARNI NG EDI TI ON. Getti ng star ted wi th the SAS
Learni ng Edi ti on. Cary, 2002. 200p.
SI LVEI RA, D.C.; MELO FI LHO, J.F.; SACRAMENTO, J.A.A.S.
& SI LVEI RA, E.C.P. Rel ao umi dade versus resi stnci a
penetrao para um Argi ssol o Amarel o di strocoeso no
recncavo da Bahi a. R. Bras. Ci . Sol o, 34:659-667, 2010.
SANTOS, H.G.; JACOMI NE, P.K.T.; ANJOS, L.H.C.;
OLI VEI RA,V.A.; OLI VEI RA, J.B.; COELHO, M.R.;
LUMBRERAS, J.F. & CUNHA, T.J.F., eds. Si stema
brasi l ei ro de cl assi fi cao de sol os. 2.ed. Ri o de Janei ro,
Embrapa Sol os, 2006. 306p.
SOI L SURVEY STAFF. Key s to s oi l tax on omy . 11.ed.
Washi ngton, DC, USDA/NRCS, 2010. 338p.
STOLF, R. Teor i a e teste exper i mental de fr mul as de
tr ansfor mao dos dados de penetr metr o de i mpacto
em resi stnci a do sol o. R. Bras. Ci . Sol o, 15:229-235, 1991.
TO, J. & KAY, B.D. Vari ati on i n penetrometer resi stance wi th
soi l pr oper ti es: The contr i buti on of effecti ve str ess and
i mpl i cati ons for pedotr ansfer functi ons. Geoder ma,
126:261-276, 2005.
TORRES, E. & SARAI VA, O.F. Camadas de i mpedi mento
mecni co do sol o em si stemas agr col as com a soj a.
Londr i na, Embr apa Soja, 1999. 58p. (Ci r cul ar Tcni ca,
23).
VAZ, C.M.P.; PRI MAVESI , O.; PATRI ZZI , V.C. & I OSSI , M.F.
I nfl unci a da umi dade na resi stnci a do sol o medi da com
pen etr metr o de i mpacto. So Car l os , Embr apa
I nstr umentao Agr opecur i a, 2002. 5p. (Comuni cado
Tcni co, 51).
VAZ, C.M.P.; MANI ERI , J.M.; MARI A, I .C. & TULLER, M.
Model i ng and cor r ecti on of soi l penetr ati on r esi stance
for var yi ng soi l water content. Geoder ma, 166:92-101,
2011.