Abstract
A survey of 78 commercial iceberg and romaine lettuce (Lactuca sativa L.) fields in the coastal valleys of central California was conducted in 2004–2005. Whole leaf samples were collected at early heading and again within 1 week of harvest. Diagnosis and Recommendation Integrated System (DRIS) leaf concentration norms were calculated for N, P, K, Ca, Mg, S, B, Zn, Mn, Fe, and Cu. Iceberg and romaine lettuce had sufficiently similar leaf nutrient concentrations that the data were combined in the DRIS calculations. Optimum leaf nutrient ranges were developed using data from high-yield fields in which all nutrients were in balance according to the DRIS approach. The DRIS-derived optimum ranges for K and Ca were substantially lower than previously published leaf sufficiency ranges, whereas for the other nutrients, the DRIS optimum ranges were in close agreement. Cu was the nutrient most frequently below the optimum range in low-yield fields. Comparison of leaf nutrient concentrations with soil nutrient availability and grower fertilization practices suggested that significant improvement in fertilizer management was possible.
The coastal valleys of central California produce more than 60,000 ha of lettuce annually, accounting for more than half of the nation's supply. Significant changes have occurred in the industry over the past decade; newer cultivars, increased planting density, and the widespread use of drip irrigation are among the factors that have increased average yields and led to modified fertilization practices. Considerable research on N and P fertilization of lettuce in this region has been conducted in recent years (Breschini and Hartz, 2002; Hartz et al., 2000; Johnstone et al., 2005). This research called into question the validity of the plant tissue nutrient sufficiency ranges currently used by the industry (Lorenz and Tyler, 1983: Ludwick, 2002).
Foliar nutrient sufficiency guidelines have often been developed from limited numbers of fertilization trials in which typically only one or two nutrients have been manipulated. That approach may not adequately reflect the influences of the wide range in soil characteristics and environmental conditions that characterize a regional industry. Furthermore, the interaction among the various macro- and micronutrients is not easily captured by traditional fertilizer trials.
An alternative approach to the development of foliar nutrient sufficiency guidelines is the Diagnosis and Recommendation Integrated System (DRIS; Beaufils, 1973). In the DRIS approach, differences in nutrient concentrations and nutrient ratios between high- and low-yielding populations are used to estimate the degree to which various nutrients may limit yield either by deficiency or excess (Walworth and Sumner, 1987). DRIS evaluation criteria have been developed for a range of agronomic and horticultural crops (Angeles et al., 1990; Beverly et al., 1984; Elwali et al., 1985; Parent and Granger, 1989; Walworth et al., 1986), including lettuce (Sanchez et al., 1991). DRIS was originally conceived as a diagnostic tool with which tissue nutrient concentrations in a field of interest could be compared with a set of established standards, or “norms”, through the calculation of nutrient indices; these indices would rank the relative degree of deficiency or excess for each nutrient. However, the complexity of this system has limited its practical application within the commercial horticultural industry.
The DRIS framework for comparing the nutrient concentration differences between high- and low-yielding populations can also be used to establish optimum leaf nutrient ranges (Needham et al., 1990). Hartz et al. (1998) used DRIS analysis to calculate leaf nutrient norms for processing tomato (Lycopersicon esculentum Mill.) and used those norms and the variance in nutrient concentrations among high-yield fields to establish leaf nutrient optimum ranges. The objective of the current study was to apply DRIS analysis to data collected in a survey of commercial lettuce fields to establish whole leaf macro- and micronutrient optimum ranges of broad applicability to the California lettuce industry.
Materials and Methods
In 2004 and 2005, a total of 35 commercial fields of iceberg and 43 fields of romaine lettuce were sampled in the coastal production regions of central California. Fields were selected to cover the growing season from early Spring through Fall and to represent a range of grower production practices. Composite soil samples (top 30 cm) were collected in each field at the start of the season, air-dried, and screened through 2-mm mesh. Soil pH was determined on saturated paste extracts and organic matter content by a modified Walkley-Black method (Nelson and Sommers, 1982). Soil texture was quantified by the hydrometer method of Sheldrick and Wang (1993). Bicarbonate extraction (Olsen and Sommers, 1982) was used to estimate soil P availability. Soil-exchangeable K, Ca, and Mg were measured by atomic emission spectrometry (AES) following ammonium acetate extraction (Thomas, 1982). DTPA-extractable soil Zn, Mn, Fe, and Cu were determined by atomic absorption spectrometry (AAS; Lindsay and Norvell, 1978).
Whole leaf samples were collected at the early heading stage (about midseason) and again within 1 week of harvest (designated the “preharvest” sample). At each growth stage, the youngest wrapper leaf was collected from at least 20 plants. Leaves were rinsed in 0.5% detergent solution to minimize surface contamination, and then oven-dried and ground for analysis. Total leaf N was determined by nitrogen gas analyzer (Model FP-528; LECO Corp., St. Joseph, Mich.). Leaf K concentration was analyzed by AES after extraction in 2% acetic acid. All other tissue analyses were conducted after microwave digestion with nitric acid and hydrogen peroxide (Sah and Miller, 1992); Zn, Mn, Fe, and Cu concentration was determined by AAS, and P, S, Ca, Mg, and B by ICP-AES.
Participating growers reported their seasonal fertilization rates and the commercial yield of each field. Additionally, they rated crop quality (good, fair, poor) and noted any field in which the yield did not reflect the productivity of the crop as a result of poor market conditions (saleable product left in the field because of low demand or price), disease, or insect damage. Based on commercial yield and grower quality ratings, the fields were divided into two groups: high-yield fields rated as “good” quality and low-yield fields rated as “poor” or “fair” quality. High yield was defined as greater than 2220 cartons/ha for iceberg lettuce and 2470 cartons/ha for romaine. Although lettuce mass per carton varies based on the type of pack, the industry average is ≈22 and 17 kg for iceberg and romaine, respectively. Low yield was defined as less than 1970 or 2220 cartons/ha for iceberg lettuce and romaine, respectively. Data from fields of intermediate yield and fields in which yield was affected by market conditions or other nonnutrient-related factors were eliminated from further use.
The mean and variance for each foliar nutrient concentration and the three possible mathematical expressions relating each pair of nutrients (i.e., N/P, P/N, N*P) were calculated for both yield groups at both growth stages. The expression that maximized the ratio low-yield variance/high-yield variance was determined for each growth stage, and the high-yield mean of that expression was selected as the DRIS norm for that nutrient pair. Nutrient pairs for which none of the mathematical expressions given an F > 1.0 were omitted from further analysis.
A DRIS index for each nutrient was calculated for each field using the method of Walworth and Sumner (1987). In short, the relative abundance of each nutrient was evaluated by comparing all expressions containing that nutrient (i.e., N/P, N*K, Ca/N, and so on) with the corresponding DRIS norms; the DRIS index is the sum of these comparisons. In theory, an index value of zero would indicate an optimum level of a nutrient, but in practice, an optimum range is more appropriate. Following the precedent of Beaufils (1973), a nutrient index within 1.33 standard deviations (sds) of the high-yield group's zero index value was considered to be balanced and sufficient for high-yield production; ±1.33 sd would encompass ≈80% of high-yield fields, assuming a normal distribution. High-yield fields were then evaluated for overall nutrient balance defined as having all nutrient indices within 1.33 sd of the group's zero index value.
Results and Discussion
A wide range of soil characteristics was encountered in the monitored fields (Table 1). In general, soils were moderately alkaline with relatively low organic matter and high P and K availability. Seasonal fertilizer application averaged 184, 19, and 26 kg·ha−1 N, P, and K, respectively, but varied widely among fields; application rates ranged from 30 to 440 kg·ha−1 N, 0–54 P, and 0–78 K. There were 29 high-yield and 25 low-yield fields. High-yield fields averaged 2370 and 3040 cartons/ha for iceberg lettuce and romaine, respectively; the low-yield average was 1460 cartons/ha for both crops. The Shapiro Wilk test confirmed that the yields of the 54 fields used in the DRIS analysis were normally distributed.
Range of initial soil characteristics (top 30-cm depth) in the fields sampled for DRIS analysis.
Minimal differences in leaf nutrient concentrations between lettuce types were found (Fig. 1). Therefore, DRIS analysis was performed on the combined iceberg and romaine data. Leaf nutrient concentrations varied between growth stages with N and P declining and K, Ca, Mg, and B increasing between the early heading and preharvest samples (Fig. 2). For most nutrient pairs, there was no ratio or product that consistently discriminated between yield groups (F > 1.0) and remained relatively numerically constant across growth stages. Therefore, separate DRIS norms were developed for each growth stage. Table 2 lists the mean leaf nutrient concentrations for high-yield fields (the DRIS norms for the nutrient/dry matter ratios); DRIS norms developed in Florida for iceberg lettuce at early heading (Sanchez et al., 1991) have also been included for comparison. The Florida norms were marginally lower for N and P, and much higher for K, Ca, and Mg.
Comparison of leaf nutrient concentrations between iceberg and romaine lettuce fields. B, Zn, Mn, Fe, and Cu in mg·kg−1; all other nutrients in g·kg−1. Bars indicate standard error.
Citation: HortScience horts 42, 1; 10.21273/HORTSCI.42.1.143
Mean leaf nutrient concentrations at early heading and preharvest growth stages; data represent all iceberg and romaine lettuce fields sampled. Bars indicate standard error.
Citation: HortScience horts 42, 1; 10.21273/HORTSCI.42.1.143
DRIS nutrient norms for lettuce: Florida norms for iceberg lettuce, California norms for iceberg and romaine.
All nutrients were in balance in 12 and 13 high-yield fields at the early heading and preharvest stages, respectively. Using only these “balanced” fields, the linear regression relationship between each DRIS nutrient index and the respective leaf nutrient concentration was calculated. All regressions were significant at P < 0.05 with the exception of N at both growth stages and B at the preharvest stage. From the significant regressions, an optimum leaf nutrient concentration range was calculated, defined as the range corresponding to the zero value of the DRIS index, ±1.33 sd of that nutrient concentration in balanced, high-yield fields. With this statistical approach, ≈80% of high-yield fields would be expected to fall within the range. For N and preharvest B, the optimum range was defined simply as the mean value ±1.33 sd of the balanced, high-yield fields.
The DRIS optimum ranges for N and P were in general agreement with previously published sufficiency ranges (Table 3). However, for K, Ca, and Mg, there was wide variation among references regarding sufficiency levels with the DRIS optimum ranges among the lowest values. The upper limit of the DRIS Ca range was below the lower limit of all other references; the highest leaf Ca encountered in this study was 11.8 and 14.1 g·kg−1 at the early heading and preharvest stages, respectively. DRIS optimum micronutrient ranges were similar to the established sufficiency ranges with the exception of Fe, which was higher in the DRIS range (Table 4). Both Hochmuth et al. (1991) and Jones et al. (1991) differentiated between iceberg and romaine sufficiency ranges, whereas in this study, the leaf nutrient concentrations were nearly identical between the two lettuce types.
Comparison of DRIS-derived leaf macronutrient optimum ranges for lettuce with published nutrient sufficiency ranges.
Comparison of DRIS-derived leaf micronutrient optimum ranges for lettuce with published nutrient sufficiency ranges.
The discrepancy between the DRIS optimum ranges developed in this study and those developed in Florida highlight the difficulty of applying tissue nutrient standards developed under dissimilar soil and environmental conditions. At early heading, none of the high-yield fields had leaf K as high as the 96.4 g·kg−1 DRIS norm reported by Sanchez et al. (1991), and less than 40% of high-yield fields had K as high as the less restrictive sufficiency threshold of 50 g·kg−1 suggested by Hochmuth et al. (1991). Similarly, not one field in the current study met the leaf Ca standards set by these other researchers. The cause of these large regional differences was not clear, particularly in light of the dominance of Ca on the soil cation exchange (on average >70%; Table 1). It is noteworthy that the relative proportions of these cations in leaves was quite similar across growth stages and locations; the molar ratio of the DRIS leaf K, Ca, and Mg norms was 1.0:1.6:0.09 in Florida (Sanchez et al., 1991) for the early heading stage, and 1.0:1.2:0.09 and 1.0:1.5:0.10 in this study for the early heading and preharvest stages, respectively.
The DRIS approach as used in this study to develop tissue sufficiency standards has both strengths and weaknesses. Using data from numerous fields selected to represent a wide range of soil, environmental conditions and grower production practices ensured that the sufficiency ranges developed have broad applicability to the regional industry. This survey approach also documented the frequency of apparent nutrient deficiency. Cu appeared to be the most commonly limiting nutrient; >40% of low-yield fields were below the Cu sufficiency range at both growth stages. Although this did not prove widespread Cu deficiency, the similarity of leaf Cu sufficiency levels among all reference sources cited in Table 4 and the fact that the mean extractable soil Cu level in these fields was far below that reported by Brown and deBoer (1983) to be representative of California soils supported that possibility. Additional research on this point is warranted. There was a statistically significant relationship between soil-extractable Cu and preharvest leaf Cu (Fig. 3). However, that relationship was insufficiently strong to set a critical soil Cu threshold.
Relationship between DTPA-extractable soil Cu and preharvest lettuce leaf Cu; regression significant at P < 0.05.
Citation: HortScience horts 42, 1; 10.21273/HORTSCI.42.1.143
A limitation of our survey approach was that without including fields of documented nutrient deficiency in the analysis, the sufficiency ranges for some nutrients may have been higher than required for optimum yield. Possible examples in this study were the N and P ranges. Widespread overfertilization in California lettuce production has been clearly documented (Breschini and Hartz, 2002; Hartz et al., 2000; Johnstone et al., 2005); these studies have shown that soil N and P availability is sufficiently high that many commercial lettuce fields require no fertilization. The likelihood of encountering yield-limiting N or P deficiency in commercial fields in this region is remote. Therefore, leaf N and P concentrations within the DRIS optimum ranges can be considered sufficient for high-yield production, but values below these ranges may not represent yield-limiting nutrient deficiency. There were no apparent adverse effects of high N or P availability on crop productivity or uptake of other nutrients; the only significant negative correlation observed was between leaf P and leaf Mg at the preharvest stage (r = −0.26), yet mean preharvest leaf Mg was the same for both yield groups (3.4 g·kg−1).
This field survey approach did provide an opportunity to evaluate the soil fertility management practices of the coastal lettuce industry. Seasonal N rate varied among fields by more than 300 kg·ha−1, but application rate was not correlated with either commercial yield or leaf N concentration. Preharvest leaf N was below the DRIS optimum range in nine fields, but only three of these were low-yielding, supporting the possibility that the N optimum range may be higher than required.
The application of P or K was more often based on grower practice than on soil fertility status (Fig. 4). Many growers used a standard fertilization program that was applied to all their fields, regardless of soil test level. That was most clearly observed with K management. Growers who routinely included K in preplant fertilization have built up high exchangeable soil K (>1.0 cmol·kg−1 in many fields); in contrast, the fields of growers who do not routinely apply K averaged <0.5 cmol·kg−1 with some fields approaching a critical level. For fields in which no K was applied, a quadratic equation described the relationship between soil K and preharvest leaf K (y = 142.8 x – 108.3 x2, r2 = 0.26). The soil K level corresponding to the lower limit of the DRIS optimum range was 0.25 cmol·kg−1.
Relationships among soil test level, grower fertilizer application, and preharvest lettuce leaf nutrient concentration for (A) P and (B) K.
Citation: HortScience horts 42, 1; 10.21273/HORTSCI.42.1.143
Similarly, the decision to apply P was often not based on soil testing with P application most common in the fields of highest soil P level. The generally high level of soil P fertility in this production region was emphasized by the lack of correlation between soil test P and leaf P at any growth stage even in fields with no applied P. These results confirm the findings of Johnstone et al. (2005) that P application in high-P soils has no effect on lettuce production.
In summary, data on soil nutrient status, leaf nutrient concentration, and grower fertilizer application rate from 78 commercial lettuce fields allowed the calculation of DRIS optimum leaf nutrient ranges and provided insight on potential improvements in fertilizer management practices. The DRIS optimum nutrient ranges generally confirmed previously published leaf sufficiency ranges for N, P, and most micronutrients. However, leaf sufficiency ranges for K, Ca, and Mg were much lower than those previously reported from Florida, highlighting the danger of using standards developed under different environmental conditions without local confirmation of their validity. An overall reduction in N application and greater reliance on soil testing to guide P and K fertilization was warranted.
Literature Cited
Angeles, D.E. , Sumner, M.E. & Barbour, N.W. 1990 Preliminary nitrogen, phosphorus and potassium DRIS norms for pineapple HortScience 25 652 655
Beaufils, E.R. 1973 Diagnosis and Recommendation Integrated System (DRIS) Soil Sci. Bull. 1 Univ. of Natal South Africa
Beverly, R.B. , Stark, J.C. , Ojala, J.C. & Embleton, T.W. 1984 Nutrient diagnosis of Valencia oranges by DRIS J. Amer. Soc. Hort. Sci. 109 649 654
Breschini, S.J. & Hartz, T.K. 2002 Presidedress soil nitrate testing reduces nitrogen fertilizer use and nitrate leaching hazard in lettuce production HortScience 37 1061 1064
Brown, A.L. & deBoer, G.J. 1983 Soil tests for zinc, iron, manganese and copper Reisenauer H.M. Soil and plant tissue testing in California Univ. Calif Bulletin 1879
Elwali, A.M.O. , Gascho, G.L. & Sumner, M.E. 1985 Sufficiency levels and DRIS norms for 11 nutrients in corn Agron. J. 77 506 508
Hartz, T.K. , Bendixen, W.E. & Wierdsma, L. 2000 The value of pre-sidedress soil nitrate testing as a nitrogen management tool in irrigated vegetable production HortScience 35 651 656
Hartz, T.K. , Miyao, E.M. & Valencia, J.G. 1998 DRIS evaluation of the nutritional status of processing tomato HortScience 33 830 832
Hochmuth, G. , Maynard, D. , Vavrina, C. & Hanlon, E. 1991 Plant tissue analysis and interpretation for vegetable crops in Florida. Univ. Florida Special Publication SS-VEC-42
Johnstone, P.R. , Hartz, T.K. , Cahn, M.D. & Johnstone, M.R. 2005 Lettuce response to phosphorus fertilization in high phosphorus soils HortScience 40 1499 1503
Jones, J.B. , Wolf, B. & Mills, H.A. 1991 Plant analysis handbook. Micro-Macro Publishing Athens, Ga 213
Lindsay, W.L. & Norvell, W.A. 1978 Development of a DTPA soil test for zinc, iron, manganese and copper Soil Sci. Soc. Amer. J. 42 421 428
Lorenz, O.A. & Tyler, K.B. 1983 Plant tissue analysis of vegetable crops 24 29 Reisenaur H.M. Soil and plant tissue testing in California. Univ. Calif. Coop. Ext Bull 1879
Ludwick A.E. 2002 Western fertilizer handbook 9th ed. Interstate Publishers, Inc Danville, Ill 356
Needham, T.D. , Burger, J.A. & Oderwald, R.G. 1990 Relationship between diagnosis and recommendation integrated system (DRIS) and optima and foliar nutrient critical levels Soil Sci. Soc. Amer. J. 54 883 886
Nelson, D.W. & Sommers, L.E. 1982 Total carbon, organic carbon and organic matter 539 579 Page A.L. , Miller R.H. & Keeney D.R. Methods of soil analysis Part 2. Chemical and microbiological properties ASA Monograph Number 9
Olsen, S.R. & Sommers, L.E. 1982 Phosphorus 403 430 Page A.L. , Miller R.H. & Keeney D.R. Methods of soil analysis: Part 2 Chemical and microbiological properties. ASA Monograph Number 9
Parent, L.E. & Granger, R.L. 1989 Derivation of DRIS norms from a high density apple orchard established in Quebec Appalachian mountains J. Amer. Soc. Hort. Sci. 114 915 919
Sah, R.N. & Miller, R.O. 1992 Spontaneous reaction for acid dissolution of biological tissues in closed vessels Anal. Chem. 64 230 233
Sanchez, C.A. , Snyder, G.H. & Burdine, H.W. 1991 DRIS evaluation of the nutritional status of crisphead lettuce HortScience 26 274 276
Sheldrick, B.H. & Wang, C. 1993 Particle size distribution 499 511 Carter M.R. Soil sampling and methods of analysis. Can. Soc. Soil Sci. Lewis Publishers Ann Arbor, Mich
Thomas, G.W. 1982 Exchangeable cations 159 165 Page A.L. , Miller R.H. & Keeney D.R. Methods of soil analysis: Part 2. Chemical and microbiological properties. ASA Monograph Number 9
Walworth, J.L. & Sumner, M.E. 1987 The diagnosis and recommendation integrated system Adv. Soil Sci. 6 149 188
Walworth, J.L. , Sumner, M.E. , Isaac, R.A. & Plank, C.O. 1986 Preliminary DRIS norms for alfalfa in the southeastern United States and a comparison with Midwestern norms Agron. J. 78 1046 1052
