0 evaluări0% au considerat acest document util (0 voturi)
12 vizualizări10 pagini
Soil penetration resistance is an important indicator of soil compaction and is strongly influenced by soil water content. This study developed mathematical models to normalize soil penetration resistance (SPR) using a reference value of gravimetric soil water content (U) pedotransfer functions were generated correlating the values of U and BD to the SPR values.
Descriere originală:
Titlu original
2012 Moraes - Correction of Resistance to Penetration by Pedofunctions and a Reference Soil Water Content
Soil penetration resistance is an important indicator of soil compaction and is strongly influenced by soil water content. This study developed mathematical models to normalize soil penetration resistance (SPR) using a reference value of gravimetric soil water content (U) pedotransfer functions were generated correlating the values of U and BD to the SPR values.
Soil penetration resistance is an important indicator of soil compaction and is strongly influenced by soil water content. This study developed mathematical models to normalize soil penetration resistance (SPR) using a reference value of gravimetric soil water content (U) pedotransfer functions were generated correlating the values of U and BD to the SPR values.
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.