i Maataloustieteellinen A ikakauskirja Vol. 63: 75—83, 1991 Phosphorus supplying capacities of soils previously fertilized with different rates of P MARKKU YLI-HALLA Kemira Oy, Espoo Research Centre, Luoteisrinne 2, SF-02270 Espoo, Finland Abstract. The residual effect of repeated P fertilizer applications was studied in a mate- rial of 30 silty clay soil samples collected from an 11-year field experiment in which a total of 0, 154, 309, 541 or 696 kg P/ha had been applied in annual doses. Half of the experiment had been limed twice with CaC03 (10 tons/ha). In a pot experiment, six yields of Italian ryegrass were grown in soils taken from each plot, and the P uptake by the grass was deter- mined. Soil P was extracted with water (P w ) and 0.5 M ammonium acetate-0.5 M acetic acid at pH 4.65 (Paaac)- Reversibly adsorbed P (Pj) was extracted by a new method in which P desorbing from the soil was collected in strips of filter paper impregnated with iron hydroxide. P uptake by pot-grown grass from soils fertilized with increasing rates of P in the field cor- responded to 30, 72, 100 and 112 kg larger quantities of P per hectare, compared to the soil not receiving P in the field experiment. The apparent utilization ofresidual fertilizer P ranged from 16 % to 25 %. The reserve of potentially desorbable P in soil had been affected much more by the fertilizer applications than had P uptake by crops in the field. The ability of the three extraction methods (P w , P iF PAAAc) to predict P uptake by pot-grown ryegrass was dis- cussed. The P ( method appeared to be well suited for assessment of potentially available P reserves both in limed and unlimed soils. Index words: water extraction, acid ammonium acetate extraction, reversibly adsorbed P, pot experiment, liming Introduction Since soil testing was started in Finland in the 19505, the average of easily soluble P, ex- tracted with an ammonium acetate solution at pH 4.65 (Paaac) (Vuorinen and Mäkitie 1955), has increased from 5.4 mg/dm3 in 1955—60 to 12.5 mg/dm3 in 1988 (Kähäri et ai. 1987, Kähäri 1989), indicating accumula- tion of soluble residual P in soils. Recommen- dations for P fertilization in Finland are based on the results of field experiments in which the relationships between yield response to P fertilization and the quantities of in the soil have been studied. Differing from fertili- zation recommendations, the actual quantity 75 JOURNAL OF AGRICULTURAL SCIENCE IN FINLAND https://www.c-info.fi/en/info/?token=86xYPJ-BnuxSkxh6.KZ-fbK0C-ePK32XQUOlkkQ.CoQYjRBodBpRerc-0xABXdiDW8qK2VxA_TCKiLAbVSe9N8ngmxcmnwjTJuRbvW0bnO1w0Wrk5KxJ3r7GRUYECuOhcrlkb6OtqguLe1D0-YR_dNQKK9hndJ33HfQ_1k4TyzdOu7SQfuU61cgpW9KDwFszSvFAaCHFElJnD9_Q of potentially desorbable P (capacity factor) cannot be determined by methods such as AAAc or water extractions, which primarily reflect the concentration of P in the soil solu- tion (intensity factor), because the relationship between the intensity and capacity factors of P is regulated, for example, by clay content and the contents of poorly crystalline Al-oxide and Fe-oxide. Neither can the reserves of potentially solubleP be estimated to be equal to the quantity of fertilizer P remaining in the soil, i.e. the difference between P fertilization and uptake by the crop, because the solubili- ty of residual P decreases over time. This phenomenon was recently exemplified in an 11-year field experiment on a silty clay soil (Yli-Halla 1989b), where an annual P ap- plication which was double the quantity of P uptake by the crop was needed to maintain the original level of P AAAc in the soil. In principle, potentially desorbable P may be solubilized in a water extraction upon in- finite solution-to-soil ratio (Madrid and Pos- ner 1979), but, in practice, procedures in- volving very wide extraction ratios are diffi- cult to perform. Strongly dissolving solutions, on the other hand, may also extract practically inert P reserves. A new method (Zee et al. 1987), based on the affinity for phosphate of freshly precipitated iron hydroxide, has been introduced for the extraction of reversibly ad- sorbed P, and may be promising for the de- termination of soluble residual P as well. In- tensive pot experiments have also been used to estimate the residual effect of P fertiliza- tion (Novais and Kamprath 1978, Steffens 1987). In the present study, the P supplying capac- ity of soils fertilized with different rates of P for 11 years was studied. The yield results of the field trial in which the current soil mate- rial originates have been published earlier (Yli-Halla 1989 b). In this paper, which is the second part of the study, the reserves of plant-available P remaining in the soil after the field experiment were determined in a pot experiment. Three extraction methods were also evaluated for their ability to differentiate between soils containing different quantities of residual P, as well as to predict P uptake by ryegrass from the respective soils. Materials and methods The soil material consisted of 30 silty clay samples taken from plough layers (A p hori- zons) of a field experiment in which the plots had been given different rates of P fertiliza- tion. Some chemical and physical characteris- tics of the soil samples, determined at the end of the field experiment, are presented in Ta- ble 1. The field experiment was conducted at Kotkaniemi experimental farm in Vihti, Southern Finland, from 1974 to 1985. The sampled plots had been fertilized annually with 0, 13/16, 26/32, 47/56 or 60/72 kg P/ha in 1974—82/1983—85. The phosphorus was given to the plots in NPK fertilizers, mainly in a water-soluble form. There were three blocks in the experiment. Half of each block Table 1. Some physical and chemical characteristics of the experimental soils. Characteristics Unlimed soils Limed soils n=lsn=ls mean range mean range Clay, % Silt, % Sand, % 31 25—40 30—37 26—45 2.1—5.1 5.8—6.1 76—91 52—86 31 35 34 3.4 6.6 77 67 25—37 35 32—36 33 24—43 Organic C, % pH(H 20) 3.4 2.4—5.5 6.4—7.1 63—91 6.0 Oxalate-extr. Fe, mmol/kg Oxalate-extr. Al, mmol/kg 8? 64 40—113 76 was limed in the spring of 1977 and of 1985 with CaC03 (10 tons/hectare). The last liming took place four months before soil sampling. The results concerning the yields of the unlimed plots have been presented in de- tail in a previous paper (Yli-Halla 1989b), in which the current trial was referred to as 'Ex- periment B'. Some data of the P status of the soil were also included. In terms of crop yields, there was no difference between the limed and unlimed plots within a given P rate. During the 11 experimental years, the total quantities of P fertilization applied, the quan- tities of P carried away in the harvested crop as well as the P balances are presented in Ta- ble 2. At the end of the field experiment, the plots were sampled and the soils were extracted with a solution containing 0.5 M ammonium acetate and 0.5 M acetic acid at pH 4.65 Table 2. Quantities of P added to the soil and taken up by the harvested crop as well as P balances of the soil at the end of an 11-year field experiment. Treatment Total P P uptake, kg/ha P balance, kg/ha kg P/ha/year fertilization “ " “ " “ ! “ ' , „,, unlimed limed unhmed limedkg P/ha 0 153 156 -153 -156 13/16 154 162 165 -8 -11 26/32 309 171 170 138 139 47/56 541 162 166 379 375 60/72 696 174 168 522 528 Table 3. Quantities of P extracted with water (Pw ) and acid ammonium oxalate (P A aac) as we" as tne quantities of reversibly adsorbed P (Pj) in soils fertilized with different quantities of P for 11 years. 1 Fertilization in the field kg P/ha/year I'» Paaac I', mg/kg mg/dm 1 mg/kg 0 unlimed 3.6 21.95.1 limed 3.8 4.9 23.9 4.5 d 4.3 d 22.9=mean 13/16 unlimed 6.7 4.7 26.1 limed 5.9 6.9 28.9 mean 6.3» 5.8' 27.5d 26/32 unlimed 9.4 5.8 33.8 limed 7.6 9.6 34.9 8.5 b 7.7 b 34.4=mean 47/56 unlimed 12.3 6.4 38.3 limed 9.1 10.9 44.3 10.4» 8.9» b 41.5 bmean 60/72 unlimed 13.7 7.7 44.7 limed 9.3 12.0 47.3 mean 11.6» 9.8» 46.0» F (P fertilization) F (liming) 48.662*** 1.965 29.997*** 23.082*** 45.727*** 0.151 Means of each level of P fertilization followed by a common letter do not differ at P = 0.05. The columns have been tested separately. *** =significant at 0.1 % level (P = 0.001) 77 78 (Paaac) (Vuorinen and Mäkitie 1955), and with deionized water (Hartikainen 1982). Further, reversibly adsorbed (P,) was ex- tracted by a method in which P desorbing from the soil is trapped by stips of filter paper impregnated with freshly precipitated iron hy- droxide (Zee et al. 1987, Yli-Halla 1989 a). All soil analyses were performed as duplicate. highest doses of P did not differfrom one an- other in terms of PAAAc and P w even though the soils proved to contain different amounts of reversibly adsorbed P (P,). In the unlimed soils, the quantities of P w were greater than those of P AAAc at all levels of previous P fer- tilization, but the opposite was observed in the limed soils. The quantities of Pi were four to five times greater than those of Pw and P AAAc . In the unlimed soils, the results of the P w, P| and P AAAc methods correlated closely with each other, but in the limed soils the results of the P AAAc method correlated only fairly with those of the P( method. The linear correlation coefficients (r) were as follows: The soils were used in a pot experiment in which six yields of Italian ryegrass (Lolium multiflorum, Lam.) were grown in 0.8 kg of soil. There were two pots from each soil in the experiment. All other nutrients except P were mixed into the soil at the beginning of the ex- periment in nutrient solutions. The com- pounds and quantities (mg/kg) used were as follows: N 300 as NH 4N03 , K 250 as KCI, Mg 38 as MgS04 • 7 H 2 O, S 50 mainly as MgS04 • 7 H2O, Zn 3as ZnS04 • 7 H2O, Cu 3as CuS04 • 5 H2O, Mn 4as MnS04 • 4 H2O, Fe 3as FeS04 • 7 H2O, B 0.5 as H 3B0 3 and Mo 0.5 as Na 2Mo04 • 2 H2O. For the succes- sive crops, N, K, Mg and S fertilizer solutions were pipetted onto the surface of the pots, the quantities of nutrients being the same as for the first yield. The sth and 6th yield were com- bined in order to get enough plant material for P analysis, performed by a vanado-mo- lybdate method (Saari and Paaso 1980). Unlimed Limed P, P\AAc Pi "aAAc P w o.96*** o.9s*** o.9l*** o.BB*** p. o.9s*** 0.70** Table 4. Quantities of dry matter produced and P taken up by ryegrass in a pot experiment. 1 Fertilization in the field kg P/ha/year Dry matter P uptake yield mg/kg g/kg of soil of soil 39.40 unlimed 36.9 limed 48.447.8 The results were studied by analyses of var- iance, followed by the Student-Neuman-Keul test (Steel and Torrie 1980) for each of the variables. Linear correlation coefficients were calculated between the results of soil analyses, yield, and P uptake by the grass. Fisher's z-transformation test (Ranta et ai. 1989) was used for testing the correlation coefficients. 42.3' 43.9 dmean 13/16 unlimed 45.0 51.4 limed 50.8 60.6 mean 47.9" 56.f> 26/32 unlimed 50.1 68.2 limed 56.0 77.3 72.8"mean 53.0» 47/56 unlimed 52.5 77.7 limed 60.0 89.6 56.2» 83.7»mean Results 60/72 unlimed 56.5 87.8 limed 60.1 89.3 The P status of the experimental soils, de- termined by the three extraction methods (Pw , P;, Paaac), reflected the different rates of P fertilizers applied to the plots in the field (Table 3). Less P was extracted by all three methods from the soils not fertilized with P in the field, compared to the fertilized soils. The soils annually receiving either of the two mean 58.3» 88.6» F (P fertilization) F (liming) 20.060'** 71.657*** 5.511* 0.6870.687 Means of each level of P fertilization followed by a com- mon letter do not differ at P=0.05. The columns have been tested separately. * =significant at 5 % level (P =0.05) *** =significant at 0.1 % level (P = 0.001) In the pot experiment, the dry matter yields of ryegrass were the higher, the more heavily the soils had been fertilized with Pin the previ- ous field experiment (Table 4). The limed soils produced, on the average, 16 "?o more dry matter than did the unlimed ones. P concen- tration of the grass, which in the first yield ranged from 1.0 to 1.7 g/kg, was also in ac- cordance with the concentration of easily solu- ble P in the soil. In the combined plant mate- rial of the sth and 6th yield, the P concen- tration of all grass samples remained below Fig. I. Uptake of P by ryegrass in the pot experiment from soils fertilized with different rates of P in the field for 11 years. The confidence limits at the 95 % level are shown. 79 1.0 g/kg, indicating a severe shortage of P. Thus, the reserves of plant-available P were practically exhausted by the end of the pot ex- periment. The differences in the P supplying capacity of the soils were clear in every ryegrass yield (Fig. 1). Total P uptake by ryegrass tops (Table 4) in the soils fertilized in the field with either of the two highest doses of P was approximately double the P uptake in the soil not receiving any P fertilization in the field. P uptake from the limed soils was, on the average, 18 % greater than that from Fig. 2. Relationships between P uptake by pot-grown ryegrass and the quantities of P extracted with water (P w) and acid ammonium acetate (P AAAc) and the quantities of reversibly adsorbed P (P,) in limed and unlimed soils. 80 the unlimed ones. In the unlimed soils, the quantities of Pw, P, and P AAAc correlated closely with the up- take of P by the plants throughout the experi- ment. In the limed soils, both the P w and the Paaac method failed to predict P uptake as accurately as they did in the unlimed soils. The relationships between the sum of P uptake by the six yields and P w , P, and P AAAc are shown in Fig. 2. In addition, the data of the limed and unlimed soils were combined and studied by correlation analyses. According to the z-transformation test, the results of the P, method were more closely (P= 1 %) corre- lated (r = o.9s***) with P uptake than were those of P w (r = o.7B***) and P AkAAc (r = o.7B***). Total P uptake by ryegrass was transformed from milligrams per kilogram of soil to kilo- grams per hectare by assuming that the plough layer of one hectare consists of 2 500 000 kg of soil. Further, the difference in P uptake in the pot experiment, owing to the fertilization treatments given in the field, was calculated as the mean of the limed and unlimed soil at each P rate. In the pot experiment, P uptake by the grass from the soils fertilized with 13/16, 26/32, 47/56 and 60/72 kg P/ha, respectively, corresponded to 30, 72, 100 and 112 kg larger quantities of P per hectare than that taken up from the soil not fertilized with P during the field experiment. Based on these differences and P balances (Table 2), it is pos- sible to estimate roughly how much of that phosphorus which made up the difference in soil P status in the field was utilized by ryegrass in the pot experiment. The following equation was used: v_ (P XD -P OD)x2 500000 xl00 "xf "of where U = utilization (%) Pxp =P uptake (mg/kg) by pot-grown rye- grass at a given P fertilization level (Ta- ble 4) Pop = P uptake (mg/kg) by pot-grown ryegrass from the soil not fertilized with P during the field experiment (Table 4) Pxf = P balance (kg/ha) at a given level of P fertilization at the end of the field ex- periment (Table 2) Pof -P balance (kg/ha) after the field experi- ment in the soil not fertilized with P (Table 2) Concerning P uptake by the pot-grown ryegrass, the differences between the soils fer- tilized with P and the unfertilized one cor- responded to 21 %, 25 %, 19 % and 16 °Io of the difference in the calculated P balances of the soils at P rates of 13/16, 26/32, 47/56 and 60/72 kg P/ha, respectively. These per- centages thus represent the apparent utiliza- tion of the difference in P status created in the field by the fertilizer applications. Discussion The pot experiment showed that the differ- ent rates of P fertilization applied during the previous field experiment had a considerable residual effect and had obviously greatly in- fluenced the pool of easily soluble P of the soil, as was also suggested by the results of the three extraction methods (P w, P,, Paaac)- P uptake by the pot-grown ryegrass cor- responded to the range of 110 to 220 kg P/ha in a plough layer consisting of 2 500 000 kg of soil, revealing the considerable reserve of potentially desorbable P in the soil. The cereal crops grown in the field experiment had tak- en up a total of 155 to 171 kg P/ha during the 11 experimental years (Yli-Halla 1989 b), and had thus responded only slightly to the wide range of plant-available P reserves in the plots even though the experimental soil was only average in P status, compared with the means of Pw , P, and presented in other studies (e.g. Hartikainen 1982, Kähä- ri 1987, Yli-Halla 1989 a). It should be pointed out that the estimates for the maxi- mum amounts of desorbable P obtained in the pot experiment must not be considered abso- lute values because the actual result of P up- take depends, for example, on the duration 81 and intensity of the pot experiment. Yet, the present results give some idea of the magni- tude of potentially desorbable P reserves con- tained in a cultivated field. According to the P balances, the soils fer- tilized annually with 26/32, 47/56 or 60/72 kg P/ha had been enriched with P during the previous field experiment. In the soil not receiving P, the reserves of P had been de- pleted, while no net change had occurred in the soil fertilized annually with 13/16 kg P/ha. In thepot experiment, apparently only 16—25 % of that phosphorus making up the difference between the soils fertilized and not fertilized with P allowed itself to be taken up by ryegrass tops, even though the plants even- tually suffered from extreme P shortage. In addition to the quantity of P in the plant tops, an undefined quantity of P was contained in ryegrass roots. However, the bulk of residual P is likely to have been immobilized in inor- ganic and organic forms unavailable to plants, because residual fertilizer P has considerably been recovered as inorganic P in the soil (Yli-Halla 1989 b). It may be supposed that recent fertilizer residues would have been more soluble than those added to the soil in the early years of the field experiment. Thus the current estimate represents the average maximum utilization of residual P accumulat- ing in the soil during 11 years. The results of the Pw, Pi and methods predicted accurately the differences in P uptake by ryegrass in the unlimed soils, and so did those of P; in the limed ones. These correlation coefficients were even higher than those obtained between the same indices and P uptake by ryegrass in a previous study (Yli-Halla 1990), in which the material con- sisted of mineral soils taken from 32 fields. The closer correlations in the present study be- tween soil analysis and plant P uptake are probably due to the more homogeneous ma- terial as far as soil properties other than P sta- tus are concerned. Increases of electrical conductivity and ex- changeable Ca are factors known to reduce the quantities of P extracted with water (Har- tikainen 1990). Probable changes of these two soil properties, attributable to liming in the current study, may be the principal rea- sons for the lower values of Pw in the limed soil, compared to the unlimed soils fertilized with the same P rate. Contrary to the change in P w, the P uptake by pot-grown grass was slightly greater from the limed plots. There- fore, the present study, in which the P supply was the primary factor limiting growth, sug- gest that the results of Pw may be vulnerable to factors other than the actual P status of the soil. On the other hand, the results of the P AAAc method were increased by liming rela- tively more than was the uptake of P by the grass, suggesting that the P AAAc method may overestimate reserves of plant-available P in recently limed soils, as has been argued by Hartikainen (1989). Information is accumulating that P> is a good indicator of the soil reserves of plant- available P, modified by P fertilization and liming. Earlier, Menon et ai. (1989), in a pot experiment, showed that the results of the P, method correlated closely with P uptake by maize from four soils fertilized with different rates of Florida rock phosphate and triple superphosphate. In the current study, this method was well able to differentiate between soils containing varying amounts of residual P which originated in easily soluble P fertiliz- ers applied to the soil over a decade. Further, the P; method was able to predict P uptake of ryegrass accurately from limed as well as from unlimed soils. The P, method may pro- vide at least a semi-quantitative measure for the reserves of plant-available P and means for studying the residual effect of P fertiliza- tion. 82 References Hartikainen, H. 1982. Water soluble phosphorus in Finnish mineral soils and its dependence on soil properties. J. Scient. Agric. Soc. Finl. 54: 89—98. 1989. Evaluation of water and ammonium acetate tests as indices for available P in limed soils. J. Agric. Sci. Finl. 61: I—6. 1990. Effect of cation species on the desorption of phosphorus in soils treated with carbonate. Z. Pflan- zenernähr. Bodenk. 152: 435—439. KähAri, J., Mäntylahti, V. & Rannikko, M. 1987. Suo- men peltojen viljavuus 1981 —1985. 105 p. Helsinki. 1989. Suomen peltojen viljavuuden kehittyminen vuo- sina 1986—1988. 11 p. Helsinki. Madrid, L. & Posner, A.M. 1979. Desorption of phos- phate from goethite. J. Soil Sci. 30: 697—708. Menon, R.G., Hammond, L.L. & Sissingh, H.A. 1989. Determination of plant-available phosphorus by the iron oxide impregnated filter paper (PJ soil test. Soil Sei. Soc. Amer. J. 53: 110—115. Novais, R. & Kamprath, E.J. 1978. Phosphorus sup- plying capacities of previously heavily fertilized soils. Soil Sei. Soc. Amer. J. 42: 931—935. Ranta, E., Rita, H. & Kouki, J. 1989. Biometria. 2nd ed. Helsinki. 569 p. Saari, E. & Paaso, A. 1980. Mineral element composi- tion of Finnish foods. 11. Analytical methods. Acta Agric. Scand. Suppl. 20: 80—89. Steel, R.G.D. & Torrie, H.J. 1980. Principles and procedures of statistics. A biometrical approach. 633 p. 2nd Ed. Singapore. Steffens, D. 1987. Einfluss langjähriger Diingung mit verschiedenen Phosphatdungerformenauf die Phos- phatverfiigbarkeit in der Rhizosphäre von Raps. Z. Pflanzenernähr. Bodenk. 150: 75—80. Vuorinen, J. & Mäkitie, O. 1955. The method of soil testing in use in Finland. Agrogeol. Publ. 63. 44 p. Yli-Halla, M. 1989 a. Reversibly adsorbed P in mineral soils of Finland. Commun. Soil Sci. Plant Anal. 20: 695—709. 1989 b. Effect of different rates of P fertilization on the yield and P statusof the soil in two long-term field experiments. J. Agric. Sci. Finl. 61: 361 —370. 1990. Comparison of a bioassay and three chemical methods for determination of plant-available P in cul- tivated soils of Finland. J. Agric. Sci. Finl. 62: 213— 219. Zee, S.E.A.T.M. van der, Fokkink, L.G.T. & Riemsdijk, W.H. van. 1987. A new technique for assessment of reversibly adsorbed phosphate. Soil Sei. Soc. Amer. J. 51: 599—604. Ms received 29.5.1990 SELOSTUS Fosforilannoituksen jälkivaikutus intensiivisessä astiakokeessa Markku Yli-Halla Kemira Oy, Espoon tutkimuskeskus, Luoteisrinne 2, 02270 Espoo Kolmenkymmenen hiuesavimaan P-tilaa tutkittiin as- tiakokeessa sekä kemiallisin analyysein. Näytteet oli otettu Vihdissä tehdystä kenttäkokeesta, jossa oli 11 vuoden ai- kana annettu NPK-lannoitteissa fosforia yhteensä 0, 154, 309, 541 tai 696kg/ha. Puolet kokeesta oli kalkittu kah- desti CaC03 :lla (10 t/ha). Astiakokeessa kasvatettiin kuusi satoa Italian raiheinää, mikä lähes ehdytti maassa olleet kasveille käyttökelpoisen P:n varat. Fosforilla lan- noitetussa maissa kasvaneen raiheinän yhteensä ottamat P-määrät vastasivat 30, 72, 100 ja 112 kg suurempia P-määriä hehtaaria kohti kuin P-otto niistä maista, joi- ta ei ollut lannoitettu fosforilla. Maahan kenttäkokeen aikana kertyneen P:n hyväksikäyttöaste vaihteli eri lan- noitustasoilla 16 %:sta 25 %:iin. Kenttäkokeessa anne- tulla P-lannoituksella oli ollut merkittävämpi vaikutus maan liukoisen P:n varoihin kuin aikanaan satojen pel- lolla ottamiin P-määriin. Viljavuusanalyysissä käytettä- vällä happamalla ammoniumasetaattiliuoksella ja vedel- lä maasta uuttuvat P-määrät selittivät kalkitsemattomis- sa maissa raiheinän P-oton vaihtelun tarkasti (R 2 = 0.94*" ja o.93***) muttakalkituissa maissa selitysasteet olivat hieman pienemmät (R 2 = o.72*** ja o.74***). Fos- forin uuttoon käytettiin lisäksi uutta menetelmää, jossa maasta liukeneva P sidotaan rautahydroksidilla kylläs- tettyyn suodatinpaperiin. Näin saadut tulokset ennusti- vat raiheinän P-oton tarkasti kalkitsemattomista (R 2 = o.94***) ja kalkituista (R 2 = o.Bs***) maista. Tällä me- netelmällä voidaan lisäksi määrittää kasveille käyttökel- poisen P:n varojakvantitatiivisemmin kuin asetaatti- tai vesiuutolla. 83