JEQ Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 7 June 2005
Published in J Environ Qual 34:1214-1223 (2005)
DOI: 10.2134/jeq2004.0393
© 2005 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Mathews, B. W.
Right arrow Articles by Tsang, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Mathews, B. W.
Right arrow Articles by Tsang, S.
Agricola
Right arrow Articles by Mathews, B. W.
Right arrow Articles by Tsang, S.
Related Collections
Right arrow Surface Water Quality
Right arrow Volcanic Soils
Right arrow Other Forage Crops
Right arrow Nutrient Management
Right arrow Phosphorus

TECHNICAL REPORTS

Plant and Environment Interactions

Phosphorus in Hawaiian Kikuyugrass Pastures and Potential Phosphorus Release to Water

B. W. Mathewsa,*, J. R. Carpenterb, L. E. Sollenbergerc and S. Tsangd

a College of Agriculture, Forestry, and Natural Resource Management, University of Hawai'i at Hilo, 200 West Kawili Street, Hilo, HI 96720-4091
b Department of Human Nutrition, Food, and Animal Sciences, College of Tropical Agriculture and Human Resources, University of Hawai'i at Manoa, 1955 East-West Road, Honolulu, HI 96822
c Agronomy Department, University of Florida, Gainesville, FL 32611-0500
d Biochemistry Department, Iowa State University, Ames, IA 50011

* Corresponding author (bmathews{at}hawaii.edu)

Received for publication October 24, 2004.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Pasture systems in Hawaii are based primarily on kikuyugrass (Pennisetum clandestinum Hochst. ex Chiov.). Relationships among kikuyugrass P concentration, animal P requirements, and various soil P determinations are needed to help identify source areas for implementing pasture management strategies to limit P loss via overland flow. A total of 51 rotationally stocked kikuyugrass pastures (>20 yr old) with contrasting soil chemical properties were sampled. A satisfactory predictive relationship between modified-Truog (MT)-extractable phosphorus (PMT) and dissolved (<0.45-µm pore diameter), molybdate-reactive phosphorus (DRP) desorbed from soil in a water extract (DRPWE) was found when 0- to 4-cm-depth data for the soil orders with medium to high DRPWE (two Mollisols and an Inceptisol) were pooled separately from those with low DRPWE (five Andisols, three Ultisols, and an Oxisol). The oxalate phosphorus saturation index (PSIox) procedure was the best predictor of DRPWE across soil orders when oxalate-extractable molybdate-reactive phosphorus (RPox) was used to calculate PSIox (PSIoxRP) rather than when total oxalate-extractable phosphorus (TPox) was used (PSIoxTP). There was little DRPWE until PSIoxRP exceeded 6% or PSIoxTP exceeded 8%. A more empirical dilute-acid phosphorus saturation index (PSIMT) was also calculated using PMT and MT-extractable iron (FeMT) and aluminum (AlMT). The PSIMT procedure showed some utility in predicting DRPWE, was positively related to the PSIox procedures, and can be more readily performed in agronomic soil testing laboratories than PSIox. The present research suggests that while Hawaiian kikuyugrass pastures tend to be sufficient to high in forage P, potential soil P release to water only appeared to be a possible environmental concern for the Mollisol and Inceptisol sites.

Abbreviations: DC, dithionite–citrate • DRP, dissolved (<0.45-µm pore diameter), molybdate-reactive phosphorus • DRPWE, dissolved, molybdate-reactive phosphorus desorbed from soil in a water extract • FeDC, AlDC, RPDC, dithionite–citrate-extractable iron, aluminum, and molybdate-reactive phosphorus, respectively • FeMT, AlMT, PMT, modified-Truog-extractable iron, aluminum, and phosphorus, respectively • Feox, Alox, RPox, and TPox, oxalate-extractable iron, aluminum, molybdate-reactive phosphorus, and total phosphorus, respectively • ICPES, inductively coupled plasma emission spectroscopy • MT, modified Truog extractant • PSIMT, a soil phosphorus saturation index estimated from modified-Truog-extractable iron, aluminum, and phosphorus • PSIMT[P/Al], a soil phosphorus saturation index estimated from modified-Truog-extractable aluminum and phosphorus only • PSIox, soil phosphorus saturation index based on oxalate extraction • PSIoxOP, extent of soil saturation with oxalate-extractable organic phosphorus • PSIoxRP, soil phosphorus saturation index calculated using oxalate-extractable iron, aluminum, and molybdate-reactive phosphorus • PSIoxTP, soil phosphorus saturation index calculated using oxalate-extractable iron, aluminum, and total phosphorus


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
KIKUYUGRASS IS THE MOST widespread pasture grass in Hawaii and is common in tropical highland and subtropical environments throughout the world (Hanna et al., 2004). This species is highly efficient in utilizing soil P and deficiencies appear to occur only in very P-infertile soils (Miles, 1997; Mathews et al., 2001). In Hawaii, minimal seasonal variation is observed within a pasture for kikuyugrass P concentration as long as soil moisture is adequate and sampling occurs at a constant growth stage (Eriksen and Whitney, 1981; Carpenter et al., 1998).

Indications are that mineral premixes for cattle (Bos spp.) in much of the world frequently contain levels of P in excess of animal requirements (Mathews et al., 2004). This may be particularly true for systems based on kikuyugrass because it is often relatively rich in P (2.7 to 4.0 g kg–1) compared with other tropical forage grasses (Hanna et al., 2004). While excessive P supplementation has generally had little negative effect on animal performance except when animals are on low Ca diets (Mathews et al., 2004), it proportionally increases concentrations of readily soluble fecal P, potentially causing environmental concerns in areas of concentrated livestock production (Mathews et al., 2004).

The overall condition of the majority of Hawaii's surface waters is relatively good. Nevertheless, contamination of surface water with P is receiving increased attention in Hawaii (Henderson and Harrigan, 2002; Laws and Ferentinos, 2003) and throughout the Australasia–Pacific region (White and Sharpley, 1996; Dougherty et al., 2004), as dissolved (<0.45-µm pore diameter), molybdate-reactive phosphorus (DRP) is often associated with water quality degradation via eutrophication.

A paucity of data exists in the tropics on the interrelationships between soil test P, P concentrations of tropical forages, and potential P desorption to water or runoff (Miles, 1997; Sotomayor-Ramírez et al., 2004). Sorn-Srivichai et al. (1988) and Miles (1997) suggested that soil test P in samples collected from the top 4 to 5 cm of pasture soil rather than more conventional sampling depths of 7.5 to 16 cm (depending on the country) may best reflect soil P effects on pasture growth. The shallower sampling depth has already been established as optimal for developing relationships between soil test P and runoff P (Torbert et al., 2002). While working with the surface 5 cm of several Inceptisols and Ultisols in the UK and Pennsylvania (USA), McDowell and Sharpley (2001) found that water-extractable, dissolved molybdate-reactive phosphorus (DRPWE) was related to DRP in surface runoff obtained from simulated rainfall. With respect to plant response, the DRPWE procedure gives no indication of a soil's capacity to rapidly replenish DRP lost to plant uptake or runoff (Negrín et al., 1996). Gusman (2004) recently demonstrated that DRPWE is of limited utility in predicting kikuyugrass P concentration in Hawaii.

It is well-established that amorphous (noncrystalline and poorly crystalline), oxalate-extractable oxides of Fe and Al are the two principle components of P retention in most noncalcareous soils (Graetz and Nair, 1999; Beauchemin and Simard, 1999). Due to differences in soil P retention capacity and therefore the extent of oxalate-extractable Fe and Al oxide saturation with oxalate-extractable P on a molar ratio basis (soil phosphorus saturation index, PSIox), researchers in North America and Europe have found that PSIox rather than P extracted by the standard agronomic soil tests was the most significant soil property influencing P desorption to water (Beauchemin and Simard, 1999; Hooda et al., 2000). Recently, more empirical, but mathematically identical P saturation indices that utilize the much smaller pool of P, Fe, and Al extracted with dilute-acid agronomic soil P test extractants such as Mehlich 1 and Mehlich 3 have been proposed as less tedious and inexpensive alternatives to PSIox (Maguire and Sims, 2002; Nair and Graetz, 2002). There has also been some interest in the use of dithionite–citrate (DC)-extractable P, Fe, and Al to predict P desorption to water in certain soils rich in microcrystalline Fe oxides (McDowell and Condron, 2001). Interestingly, the procedures discussed above to predict DRPWE or DRP in surface runoff have received relatively limited attention in the tropics (Sotomayor-Ramírez et al., 2004).

The principle objectives of the present study were to (i) examine for a range of Hawaiian soils the relationship between kikuyugrass P concentration and modified-Truog (MT)-extractable phosphorus (PMT, the standard dilute-acid agronomic P test in Hawaii) for the 0- to 4- and 0- to 16-cm soil depths; (ii) determine the approximate PMT values where P supplementation of grazing cattle could be reduced or perhaps eliminated during nondrought periods; and (iii) examine for the 0- to 4-cm soil depth the relationships among DRPWE, PMT, PSIox calculated using oxalate-extractable, molybdate-reactive phosphorus (PSIoxRP), and total oxalate-extractable phosphorus (PSIoxTP); and possible alternative P saturation indices determined with the MT dilute-acid (PSIMT) and DC soil tests.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Pasture Sites and Sampling
During July and August 2002 soil and forage samples were collected from 12 representative kikuyugrass pasture sites in Hawaii (19–22° N) (Table 1) with contrasting soil chemical properties (Table 2). The soil series included five Andisols (Ainakea, Hanipoe, Kapapala, Maile, and Palapalai), an Inceptisol (Kohala), two Mollisols (Hawi and Waialua), three Ultisols (Halawa, Kalae, and Makawao), and an Oxisol (Hanamaulu–Puhi intergrade). The very high to extremely high total P concentrations in two of the soils from the slopes of the Kohala Mountains (Kohala and Palapalai) compared with other soils of Hawaii have been noted previously (Vitousek et al., 2004). All pastures sampled were >20 yr old and rotationally stocked with cattle throughout the year at a rate of 1000 ± 500 kg liveweight ha–1 (1–4 head ha–1). Grazing periods averaged 1 to 2 wk and were followed by 3- to 5-wk rest periods for regrowth from a residual stubble height of approximately 5 to 7 cm. At one site (Maile), 18 pastures were sampled that were used in several long-term research trials (Carpenter et al., 1998; Mathews et al., 2001) and ranged from very low to high in P fertility due to differential P fertilization. Three pastures each were sampled at the other "producer" sites. At some producer sites there was substantial variability among pastures with respect to physical features such as slope while others were nearly uniform.


View this table:
[in this window]
[in a new window]
  		Table 1. Site history and classification of the soils from 12 long-term kikuyugrass pasture sites.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Selected average chemical characteristics of the pasture soils studied (0- to 16-cm depth).

 
Soils were sampled at the 0- to 4-, 4- to 16-, and 0- to 16-cm soil depths by collecting a well-mixed composite of 10 cores (2-cm diameter) from the bulk area of each pasture while avoiding dung and urine patches. The composited samples for each pasture and depth were sieved through a 4-mm screen. Kikuyugrass regrowth (green leaf and associated stem) was hand-plucked to a 5-cm stubble within 30 cm of each soil sampling point and composited for each pasture. All pastures were sampled at the five-to-six-leaf-per-tiller stage of development (corresponds to 3 to 4 wk regrowth), the point when kikuyugrass nutritive value is expected to be near optimal (Reeves et al., 1996; Hanna et al., 2004).

Soil and Plant Analysis
The soil samples were maintained field moist for analysis and results are reported on an oven-dry (100°C) soil equivalent basis. All samples were analyzed for MT (0.01 M H2SO4 + 0.02 M [NH4]2SO4)-extractable phosphorus (PMT) by the molybdenum-blue colorimetric method routinely used for this soil test (Hue et al., 2000). The MT extracts from the 0- to 4-cm depth were also analyzed for extractable iron (FeMT) and aluminum (AlMT) by inductively coupled plasma emission spectroscopy (ICPES) (Hue et al., 2000). Phosphorus that is nearly immediately bioavailable or could be easily desorbed from the surface 4 to 5 cm of soil and redistributed into runoff during an intense rainstorm is thought to be represented by DRPWE (Negrín et al., 1996; McDowell and Sharpley, 2001). The DRPWE was determined for the 0- to 4- and 4- to 16-cm depths using a soil-to-solution ratio of 1:5 and a shaking time of 30 min as outlined by McDowell and Sharpley (2001), except that a correction was made when the extract was slightly colored by organic substances to account for absorption that was not due to phosphomolybdenum blue during the colorimetric analysis. This was achieved by performing an analysis of an aliquot of each extract without addition of ascorbic acid to the molybdate reagent (Negrín et al., 1996). Soil-organic C (modified Walkley–Black) and pH (saturated paste) were determined for the 0- to 4- and 4- to 16-cm-depth soil samples by the methods of Hue et al. (2000).

An acid (pH 3) Na oxalate (0.2 M) solution was prepared immediately before soil extraction by dissolving 10.92 g oxalic acid and 15.19 g Na oxalate L–1 deionized H2O (Parfitt, 1989). The oxalate-extractable reactive phosphorus (RPox), total phosphorus (TPox), iron (Feox), and aluminum (Alox) were determined for soil samples from the 0- to 4-cm depth using a soil-to-solution ratio of 1:100 and a shaking time of 4 h in the dark as recommended by Parfitt (1989) for soils derived from volcanic materials. The potential for oxalate interference during the colorimetric RPox analysis was overcome by aliquot dilution (50- to 100-fold) and the addition of excess molybdate (He et al., 1998) while TPox, Feox, and Alox were determined by ICPES (Hooda et al., 2000). The color in the oxalate extracts from the Andisols (Parfitt, 1989) did not interfere with RPox analysis due to the high dilution. The PSIox of each sample was calculated as PSIoxRP and PSIoxTP using RPox or TPox concentration (mmol kg–1) divided by the sum of Feox and Alox concentrations (mmol kg–1), and multiplied by 100 (Hooda et al., 2000). The difference between PSIoxTP and PSIoxRP was defined as the extent of soil saturation with oxalate-extractable organic phosphorus (PSIoxOP). The PSIMT for the 0- to 4-cm soil depth was calculated in the same manner as PSIoxRP, but with the substitution of PMT, FeMT, and AlMT.

Dithionite–citrate (DC) extracts of iron (FeDC), aluminum (AlDC), and RP (RPDC) were also performed for the 0- to 4-cm depth to provide an estimate of soil Fe, Al, and RP associated with crystalline Fe minerals and Al-substituted goethite–hematite in addition to the amorphous Feox and Alox minerals (Ross and Wang, 1993). The caveat in the case of the Andisols is that DC is often less than 50% as effective as oxalate in dissolving allophane (Ross and Wang, 1993). In some instances the allophane not dissolved by DC may occlude Fe that would otherwise be dissolved thereby resulting in lower levels of FeDC than Feox (Dahlgren et al., 2004). The DC (0.8 g Na dithionite in 50 mL of 0.68 M Na citrate) extractions were performed using a soil-to-solution ratio of 1:100 as described by Ross and Wang (1993). The potential for citrate interference during the colorimetric analysis of RPDC was eliminated by the procedures of He et al. (1998). The FeDC and AlDC in the DC extracts were determined by ICPES following appropriate dilution.

The hand-plucked kikuyugrass samples were dried at 60°C for 48 h and ground to pass a 1-mm stainless steel screen, using a Wiley mill. These samples were acid-digested and prepared for P and Ca analysis by ICPES as outlined by Hue et al. (2000).

All statistical relationships were determined by correlation and linear and curvilinear (e.g., exponential rise to maximum, power function) regression analyses using SigmaPlot (SPSS, 2000).


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Soil Test Phosphorus and Kikuyugrass Phosphorus and Calcium
The PMT, DRPWE, and soil organic C concentrations for the 0- to 4-cm depth were greater than those observed for the 4- to 16-cm depth (Tables 3 and 4). However, there was a trend in the opposite direction for pH in seven of the soils. This may have been due to acidity contributed by dissociation of organic matter functional groups and greater microbial activity in the surface few centimeters of soil (Loganathan and Swindale, 1969). According to McLaughlin et al. (1990) significant stratification of soil test P and soil organic C may develop within 7 to 8 yr after pasture establishment. This was attributed to pastures being no-till systems where most of the recycled and fertilizer P does not readily leach more than a few centimeters, except in sandy or cracking soils with very low P retention capacities. While the present study was conducted under conditions of adequate soil moisture, some heavily grazed pastures may be susceptible to P deficiencies if the surface few centimeters of relatively unshaded soil dry out (McLaughlin et al., 1990). This would be expected to be most problematic for soils with little capacity to maintain significant DRPWE or otherwise supply P to roots at lower soil depths during droughts (Mathews et al., 2004).


View this table:
[in this window]
[in a new window]
  		Table 3. Effect of soil on modified-Truog soil test phosphorus (PMT) concentrations at the different sampling depths, and concentrations and ratios of P and Ca in kikuyugrass. Data are means (n = 18 for the Maile soil while n = 3 for the other soils) with standard deviations in parentheses.

 

View this table:
[in this window]
[in a new window]
  		Table 4. Soil pH, water-extractable, dissolved (<0.45-µm pore diameter), molybdate-reactive phosphorus (DRPWE), and soil organic carbon (OC) for the 0- to 4- and 4- to 16-cm sampling depths of the pasture soils. Data are means (n = 18 for the Maile soil while n = 3 for the other soils) with standard deviations in parentheses.

 
Critical agronomic soil-P test concentrations have not been well defined for kikuyugrass growth but they are generally thought to be in the range of 10 to 25 mg kg–1 depending on the extraction procedure and defoliation management (e.g., grazing vs. intensive mechanical harvesting) (Miles, 1997; Mathews et al., 2001, 2004). The relationship between PMT and kikuyugrass P across soils with the exclusion of the six Mollisol outliers (denoted by open circles in Fig. 1) indicates that kikuyugrass P concentrations in grazed pastures at the five-to-six-leaf-per-tiller growth stage will tend to be deficient from the standpoint of plant (<2.2 g P kg–1 dry matter; Andrew and Robbins, 1971) and ruminant nutrition (<2.6 g P kg–1 dry matter; Mathews et al., 2004) when PMT concentrations in the 0- to 16-cm depth are <9 and <18 mg kg–1, respectively. Most of the variability within this relationship (Fig. 1) was due to variation among the Andisols (kikuyugrass P in the Andisols = 1.22 + 2.03[1 – exp(–0.07PMT)], R2 = 0.54, significant at the 0.001 probability level, n = 30) as the relationship for the pooled Inceptisol, Ultisol, and Oxisol data was much better (kikuyugrass P = 2.14 + 1.70[1 – exp(–0.02PMT)], R2 = 0.80, significant at the 0.001 probability level, n = 15) as was the relationship within the Maile soil (Andisol) alone (kikuyugrass P = 1.87 + 0.03PMT, R2 = 0.81, significant at the 0.001 probability level, n = 18). The variability among Andisols has long been noted for P (Fox and Searle, 1978) and reflects the wide ranges in mineral weathering found within Hawaiian Andisols (Loganathan and Swindale, 1969). The extremely high kikuyugrass P concentrations for the Mollisol outliers can be attributed to the very high PMT (Table 3) and DRPWE (Table 4) concentrations. These Haplustolls are known to have a very low ability to retain P compared with most other Hawaiian soils (Fox and Searle, 1978). The Waialua soil had formerly received dairy manure applications (Table 1) and had PMT concentrations two- to threefold greater than nearby abandoned row crop (mainly sugarcane, Saccharum spp.) lands on this soil (Mathews, unpublished data, 1999, 2001). Previous research has shown that the P concentrations of tropical grasses can readily exceed the typical range of 1.5 to 4.0 g P kg–1 in heavily manured soils or certain other high P soils with relatively low P sorption capacities (Mathews et al., 2004). The lack of greater sensitivity of standard agronomic soil P tests (PMT, Olsen, and Bray P-1) to the P status of tropical grasses across Hawaii's many soils has also long been recognized (Clements, 1980; Gusman, 2004). However, an efficient alternative test has yet to be developed (Gusman, 2004). With respect to soil sampling depth, the relationship between kikuyugrass P and PMT for the 0- to 4-cm depth exclusive of the Mollisol outliers (kikuyugrass P = 1.55PMT(0.16), R2 = 0.58, significant at the 0.001 probability level, n = 45) was similar to but not superior to that obtained for the 0- to 16-cm depth.



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 1. Relationship between kikuyugrass P concentration and modified-Truog-extractable phosphorus (PMT) for the 0- to 16-cm sampling depth exclusive of the Mollisol outliers (denoted by open circles). ***Significant at the 0.001 probability level.

 
While kikuyugrass Ca concentrations (Table 3) were sufficient from the standpoint of grass nutrition (≥1.5 g kg–1; Awad et al., 1976), they were deficient for grazing ruminants (<3.0 g kg–1; Mathews et al., 2004) on all soils except Maile. This occurred despite relatively high soil Ca concentrations (Table 2) and may reflect the young growth stage of the kikuyugrass (Reeves et al., 1996). The animal nutrition concern with low forage Ca is magnified by kikuyugrass Ca to P ratios that were below the recommended range of 1.0 to 2.0 for grazing livestock (Hanna et al., 2004; Mathews et al., 2004) at nine of the sites (Table 3) and because most of the Ca in kikuyugrass is thought to be present as Ca oxalate (not measured) (Gusman, 2004; Mathews et al., 2004). Both of these factors are known to contribute to reduced Ca absorption by cattle and particularly horses (Equus caballus), leading to bone and muscular-skeletal disorders (Gusman, 2004; Mathews et al., 2004). Furthermore, excessive P supplementation may further confound Ca to P imbalance (Mathews et al., 2004).

The kikuyugrass P and Ca data in the present and other Hawaiian studies (Eriksen and Whitney, 1981; Carpenter et al., 1998; Mathews et al., 2001; Gusman, 2004, and studies cited therein) provide little justification for the standard use by the Hawaiian cattle industry of free choice mineral supplement premixes containing 80 to 100 g P kg–1 in kikuyugrass pasture systems. The recommendation for this supplementation level was based on research conducted in the 1950s on rangelands in the continental United States and Hawaii research with excessively mature ‘Pangola’ digitgrass (Digitaria eriantha Steud.) (1.4 g P kg–1 dry matter; Campbell et al., 1971). More recently, Espinoza et al. (1991b) demonstrated that a free-choice mineral supplement containing 60 g P kg–1 was sufficient for cow–calf operations on bahiagrass (Paspalum notatum Flügge) pastures in central Florida that averaged much lower in forage P (1.8 g P kg–1; Espinoza et al., 1991a) than typically observed for kikuyugrass (Hanna et al., 2004). A recent private industry report recommended reducing the P concentration in mineral premixes for Hawaiian kikuyugrass pastures to 50 to 60 g kg–1 while increasing the premix Ca to P ratio from 1:1 to 2.5:1 (Dr. Russ Wyllie, Zinpro Corporation, Reno, NV, unpublished mineral supplementation strategies report prepared for Dr. H.M. "Tim" Richards, D.V.M., Veterinary Associates, Kamuela, HI, July 2004).

Prediction of DRPWE in the Surface Four Centimeters of Soil
Beauchemin and Simard (1999) and Hooda et al. (2000) suggested that a single combined relationship between P determined by any of the common agronomic soil P tests and DRPWE would be of little utility in predicting DRPWE across a wide range of soils. This was attributed to differences in soil P retention capacity or PSIox at similar soil test P concentrations. The PMT and DRPWE data in Fig. 2 for the full range of Hawaiian soils at the 0- to 4-cm depth support this suggestion. However, Fig. 2 shows that a better relationship was found when the data for the soils with medium to high DRPWE (Mollisols and the Inceptisol) were pooled separately from the soils with low DRPWE (Andisols, Ultisols, and the Oxisol). It is also worth noting that additional data (Mathews, unpublished, 2004) were collected in 2004 from a 25-yr-old mixed kikuyugrass–guineagrass [Panicum maximum (Jacq.) L.] pasture site on a Kohala series soil that had suffered fertility decline under sugarcane cultivation. These data indicated that the relationship for the Mollisols and the Inceptisol could be reliably extended downward to much lower PMT (49 ± 4 mg kg–1) and DRPWE (1.56 ± 0.24 mg L–1) concentrations.



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 2. Relationship between water-extractable, dissolved (<0.45-µm pore diameter), molybdate-reactive soil phosphorus (DRPWE) and modified-Truog-extractable phosphorus (PMT) for the 0- to 4-cm sampling depth. ***Significant at the 0.001 probability level.

 
It has often been suggested that the PSIox procedure or possibly a similar index based on a dilute-acid soil extraction procedure would be a good indicator of DRPWE across soil types (Beauchemin and Simard, 1999; Hooda et al., 2000; Maguire and Sims, 2002; Nair and Graetz, 2002). The effect of soil on mean mmol kg–1 Feox, FeMT, Alox, AlMT, RPox, PMT, and TPox for the 0- to 4-cm sampling depth are presented in Table 5 and the resulting PSIoxRP, PSIoxTP, and PSIMT values are presented in Table 6. The PSIoxRP was an improvement to PMT in predicting DRPWE across all the soils (Fig. 3) and little DRPWE was predicted until PSIoxRP exceeded 6%. The PSIoxTP has often been used for convenience in Europe and North America because Feox, Alox, and TPox can be simultaneously determined by ICPES without a separate colorimetric RPox analysis. In the present study, however, PSIoxTP proved to be a slightly weaker parameter for the estimation of DRPWE than PSIoxRP (DRPWE = 0.003PSIoxTP(2.98), R2 = 0.92, significant at the 0.001 probability level, n = 51). This can be attributed to the fact that PSIoxTP includes variable amounts of organic P depending on the soil (Mallarino and Murrell, 2004) and PSIoxOP was negatively correlated with DRPWE (r = –0.28, significant at the 0.05 probability level). Because of the inclusion of PSIoxOP in PSIoxTP, little DRPWE was predicted until PSIoxTP exceeded 8%. This is lower than the critical value of 10% PSIoxTP observed by Hooda et al. (2000) and may have resulted from the twofold greater soil organic C concentrations in the present study (Table 4). Mineralization of organic matter can serve as a dynamic source of DRPWE (He et al., 2003; Dougherty et al., 2004), and the binding of humic materials and low molecular weight organic acids to oxide surfaces has been shown to decrease phosphate adsorption per mol of Fe and Al (Easterwood and Sartain, 1990; Hue, 1991). Another contributing factor may have been greater extraction of Fe and Al by the oxalate extraction procedure utilized in the present study as discussed by Parfitt (1989) and Comfort et al. (1991).


View this table:
[in this window]
[in a new window]
  		Table 5. Effect of soil on modified-Truog-extractable iron (FeMT), aluminum (AlMT), and phosphorus (PMT); oxalate-extractable iron (Feox), aluminum (Alox), reactive phosphorus (RPox), and total phosphorus (TPox); and dithionite–citrate-extractable iron (FeDC), aluminum (AlDC), and reactive phosphorus (RPDC) for the 0- to 4-cm sampling depth. Data are means (n = 18 for the Maile soil while n = 3 for the other soils) with standard deviations in parentheses.

 

View this table:
[in this window]
[in a new window]
  		Table 6. Soil P saturation indices calculated using oxalate-extractable iron (Feox), aluminum (Alox), and molybdate-reactive phosphorus (RPox) (PSIoxRP); Feox, Alox, and total oxalate-extractable phosphorus (TPox) (PSIoxTP); and modified-Truog-extractable iron (FeMT), aluminum (AlMT), and phosphorus (PMT) (PSIMT). Data are means (n = 18 for the Maile soil while n = 3 for the other soils) with standard deviations in parentheses.

 


View larger version (16K):
[in this window]
[in a new window]
  		Fig. 3. Relationship between water-extractable, dissolved (<0.45-µm pore diameter), molybdate-reactive soil phosphorus (DRPWE) and the soil phosphorus saturation index calculated using oxalate-extractable iron, aluminum, and molybdate-reactive phosphorus (PSIoxRP). ***Significant at the 0.001 probability level.

 
The PSIMT procedure showed some utility in predicting DRPWE (Fig. 4) and was positively related to PSIoxRP (Fig. 5). A similar, but slightly poorer, relationship was found between PSIMT and PSIoxTP (PSIMT = 0.027PSIoxTP(2.63), R2 = 0.90, significant at the 0.001 probability level, n = 51). Again, this may have been because the colorimetric procedure for determining PMT includes negligible organic P while PSIoxTP includes variable amounts of organic P depending on the soil (Mallarino and Murrell, 2004). The power rather than linear regression relationships between PSIMT and the PSIox procedures suggests that the ability of MT to extract Fe or Al relative to oxalate decreased, particularly at greater values of PSIoxRP and PSIoxTP (Fig. 5). As expected based on research with other dilute acid soil test extractants (Wang et al., 1991; Maguire and Sims, 2002), a potential limitation of PSIMT is that MT is a very weak extractant for Fe compared with Al (Table 5). When averaged for all soil samples, the mean FeMT to mean Feox ratio was only 0.004 and FeMT made up only 7 ± 12% of mean FeMT + AlMT (mmol kg–1) in the PSIMT calculation. Furthermore, FeMT was not correlated with Feox (r = 0.13, P = 0.35). The ratio of mean AlMT to Alox was 0.10 and AlMT and Alox were related (AlMT = 0.09Alox + 18.08, R2 = 0.77, significant at the 0.001 probability level). However, a much poorer relationship was obtained between PMT and RPox (PMT = 0.024RPox + 1.62, R2 = 0.31, significant at the 0.001 probability level) and the ratio of mean PMT to Pox was 0.03. The overall importance of Fe in the calculation of PSIMT is demonstrated by the observation that when Fe is removed from the denominator (PSIMT[P/Al] = 100 x mmol PMT kg–1/mmol AlMT kg–1) the R2 for the relationship to DRPWE is reduced from 0.93, significant at the 0.001 probability level (Fig. 4), to 0.75, significant at the 0.001 probability level (DRPWE = 0.16PSIMT[P/Al] + 0.16, n = 51).



View larger version (16K):
[in this window]
[in a new window]
  		Fig. 4. Relationship between water-extractable, dissolved (<0.45-µm pore diameter), molybdate-reactive soil phosphorus (DRPWE) and a soil phosphorus saturation index estimated from modified-Truog-extractable iron, aluminum, and phosphorus (PSIMT). ***Significant at the 0.001 probability level.

 


View larger version (16K):
[in this window]
[in a new window]
  		Fig. 5. Relationship between a soil phosphorus saturation index estimated from modified-Truog-extractable iron, aluminum, and phosphorus (PSIMT) and the soil phosphorus saturation index calculated using oxalate-extractable iron, aluminum, and molybdate-reactive phosphorus (PSIoxRP). ***Significant at the 0.001 probability level.

 
As expected, FeDC was greater than Feox in all soils but two of the Andisols (Kapapala and Palapalai, Table 5) where allophane not dissolved by DC may have occluded Fe that otherwise may have been dissolved by this extractant (Dahlgren et al., 2004). The FeDC was 1.2- to 1.5-fold greater than Feox for the other Andisols, 3.2- to 3.6-fold greater for the Inceptisol and the Mollisols, and 3.4- to 14-fold greater for the Ultisols and the Oxisol. The AlDC concentration in the Andisols was always about half to slightly more than half that of Alox, reflecting the lower extractability of Al from allophane by DC (Ross and Wang, 1993). The ratio of AlDC to Alox was within the range of 0.9 to 1.4 for the remaining soils with three exceptions. The concentration of AlDC was twice that of Alox for the Halawa series (Ultisol) and nearly fivefold greater than Alox for the Hanamaulu–Puhi Oxisol. Greater concentrations of AlDC than Alox have been attributed to Al extraction from Al-substituted Fe minerals (Ross and Wang, 1993). Considering the mineralogy data available for the soils the source was probably Al-substituted goethite (USDA Soil Conservation Service, 1976). Interestingly, the ratio of AlDC to Alox was 0.65 for the Makawao series, perhaps reflecting that this Ultisol that had weathered from volcanic ash still contained substantial allophane.

The ratio of RPDC to RPox was between 0.8 to 1.1 for the Andisols, 1.2 to 1.6 for Mollisols and the Inceptisol, 1.8 to 16.0 for the Ultisols, and 21.5 for the Oxisol. Ratios less than unity for two of the Andisols (Hanipoe and Palapalai) were most likely due to P bound to allophane not dissolved by DC (Ross and Wang, 1993). Phosphorus fractionation of these soils by a modified Hedley procedure (Graetz and Nair, 1999) indicated only nil to trace amounts of Ca-P (Mathews, unpublished data, 2002). Therefore, it is doubtful that ratios less than unity resulted from Ca-P forms (i.e., apatite-type minerals) that could be partially dissolved by acid oxalate but not DC (Graetz and Nair, 1999). Ratios of RPDC to RPox greater than unity for all the other soils can be attributed to additional P dissolved from crystalline Fe minerals (Ross and Wang, 1993; Graetz and Nair, 1999). These ratios indicate that, with the exception of the Makawao series where only 44% of the RP appears to be associated with crystalline Fe minerals, the remaining Ultisols and the Oxisol had 82 to 95% of their RP associated with crystalline Fe minerals. Phosphorus associated with crystalline Fe minerals is relatively stable with the only possibility for significant desorption occurring under extended waterlogged conditions (Parfitt, 1979). The ratio of RPDC to FeDC on a molar basis was weakly correlated to DRPWE (r = 0.25, P = 0.08). Better correlations were obtained between DRPWE and the molar ratio of RPDC to AlDC (r = 0.73, significant at the 0.001 probability level) and the molar ratio of RPDC to the sum of FeDC and AlDC (r = 0.52, significant at the 0.001 probability level). However, none of these correlations suggest particularly strong relationships to DRPWE compared with those obtained for the PSIoxRP and PSIMT procedures.

Potential Environmental Implications of DRPWE
Using DRPWE concentration in the surface few centimeters of soil to estimate runoff DRP is not simple. This is because runoff DRP concentration can vary greatly depending on rainfall intensity and hence differing runoff flow rates, soil–water contact or residence times, water-to-soil ratios, and other flow conditions (Laws and Ferentinos, 2003; Dougherty et al., 2004). A host of landscape and biological factors such as plant biomass can also influence the various relationships (Dougherty et al., 2004). While working with the Waimanalo agricultural watershed in Hawaii, Laws and Ferentinos (2003) found that post-rain stream DRP concentrations decreased twofold and base flow DRP concentrations decreased 10- to 20-fold compared with concentrations observed during storms. They attributed this observation in part to increased reaction of DRP with soil and sediment Fe during periods of slower flow.

Suggested target runoff DRP concentrations of <0.10 mg L–1 have been proposed for grazed watersheds in the Australasia–Pacific region and may or may not be practical depending on factors such as the magnitude of transport (flow volume), dilution, and sorption downstream and the properties and use of the receiving water body (White and Sharpley, 1996; Dougherty et al., 2004). Limited runoff data collected by Mathews (unpublished, 2000, 2002) under conditions of intense natural rainfall in fall 2000 and spring 2002 for the Maile, Hawi, and Kohala soils indicated that DRPWE for the 0- to 4-cm depth averaged approximately 10 ± threefold greater than runoff DRP concentration (mg L–1). Therefore, the relatively high DRPWE concentrations observed for the 0- to 4-cm depth of the Hawi (mean = 8.87 mg kg–1) and Kohala (mean = 4.14 mg kg–1) pasture soil sites (Table 4) could be of potential environmental concern under such conditions. The Waialua pasture soil also had a relatively high DRPWE concentration (mean = 6.58 mg kg–1) for the 0- to 4-cm depth (Table 4) and there have already been concerns regarding runoff and the conditions in adjacent marshlands. However, the operation is no longer applying dairy manure to the pastures, has taken other steps to minimize P inputs and runoff, and may close in the near future. While P inputs to the Waialua site would cease with abandonment of the operation, the effect of previous inputs can last for years (Graetz and Nair, 1999; Nair and Graetz, 2002). It is also worth noting that the Kohala Mountain pasture soil sites (Hawi, Kohala, and Palapalai) are often naturally high in readily extractable P (Vitousek et al., 2004) indicating that any proposed environmental regulatory criteria for P in Hawaiian pastures should not be guided solely on the establishment of some uniform soil test P, DRPWE, or runoff DRP concentration.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Results of the present research agreed with studies conducted in North America and Europe indicating that the PSIox procedures, especially PSIoxRP, could be useful aids in estimating the potential for P losses from Hawaiian soils to surface waters. The less tedious and more routinely applicable PSIMT procedure showed some utility in predicting DRPWE and was related to the PSIox procedures. Including FeMT and AlMT analyses and the calculation of PSIMT would be a relatively small extra step for soil testing laboratories already determining PMT as Hawaii's standard agronomic P test. While this option should be further investigated, a limitation to the use of the PSIMT procedure to predict DRPWE appeared to be that MT is a very weak extractant for Fe. A satisfactory predictive relationship between PMT and DRPWE was found when data for 0- to 4-cm depth samples from the soil orders with high DRPWE (Mollisols and the Inceptisol) were pooled separately from the soil orders with low DRPWE (Andisols, Ultisols, and the Oxisol).

Kikuyugrass P concentrations in grazed pastures at the five-to-six-leaf-per-tiller stage of development tended to be deficient from the standpoint of plant and ruminant nutrition when PMT concentrations in the 0- to 16-cm depth were <9 and <18 mg kg–1, respectively. Soil samples from the 0- to 4-cm depth did not result in a superior predictive relationship between PMT and kikuyugrass P than those from 0- to 16-cm depth. Sufficient to high kikuyugrass P concentrations coupled with low Ca concentrations resulted in Ca to P ratios considerably below the recommended range of 1.0 to 2.0 for grazing livestock at 75% of the sites despite relatively high concentrations of soil Ca. Two sites were as low as 0.4. The present research indicates that while Hawaiian kikuyugrass pastures tend to be sufficient to high in forage P in terms of both plant and animal nutrition, potential soil P release to water is not likely to be a priority environmental concern for most sites. Regardless, research should be conducted to assess the potential of reducing costly P supplementation of cattle on Hawaiian kikuyugrass pastures by at least 25 to 50% and increasing Ca supplementation.


   ACKNOWLEDGMENTS
 
Funding for this research was provided by the County of Hawaii, Department of Research and Development; and the University of Hawai'i at Hilo's Hutchinson Fund for Applied Agricultural Research.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2005 34: v. [Full Text]  



This article has been cited by other articles:


Home page
 Crop Sci.Home page
J. C. B. Dubeux Jr., L. E. Sollenberger, B. W. Mathews, J. M. Scholberg, and H. Q. Santos
Nutrient Cycling in Warm-Climate Grasslands
Crop Sci., May 31, 2007; 47(3): 915 - 928.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Mathews, B. W.
Right arrow Articles by Tsang, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Mathews, B. W.
Right arrow Articles by Tsang, S.
Agricola
Right arrow Articles by Mathews, B. W.
Right arrow Articles by Tsang, S.
Related Collections
Right arrow Surface Water Quality
Right arrow Volcanic Soils
Right arrow Other Forage Crops
Right arrow Nutrient Management
Right arrow Phosphorus

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals