Silicon transfers in a rice field in Camargue (France)

Journal of Geochemical Exploration 88 (2006) 190 – 193 www.elsevier.com/locate/jgeoexp Silicon transfers in a rice field in Camargue (France) V. Desplanques a,b, L. Cary b, J.-C. Mouret c, F. Trolard b, G. Bourrié b, O. Grauby d, J.-D. Meunier a,* a CNRS/Université Paul Cézanne, CEREGE, BP 80, F13545 Aix-en-Provence Cedex, France INRA UR Géochimie des Sols et des Eaux, BP 80, F13545 Aix-en-Provence Cedex 4, France c INRA–ENSAM–IRD UMR Innovation, 2 place Viala, F 34060 Montpellier, France Université Paul Cézanne, CNRS, CRMCN, Campus de Luminy, F13288 Marseille Cedex 9, France b d Received 29 March 2005; accepted 19 August 2005 Available online 21 November 2005 Abstract We conducted a study of the biogeochemical cycle of silicon in a rice field in Camargue (France) in order to evaluate the role of biogenic silicon particles (BSi) in the cycle. Opal-A biogenic particles (phytoliths, diatoms. . .), which dissolve more rapidly than other forms of silicate usually present in soils, are postulated to represent the easiest bioavailable Si for rice. We found 0.03–0.06 wt.% of BSi in soils (mainly phytoliths). This value is lower than other values from the literature. Each year, the exportation of BSi from rice cultivation is 270 F 80 kg Si ha 1. We show that BSi input by irrigation is mostly composed of diatoms and we estimate it at 100 kg Si ha 1 year 1. This value is more than a third of the annual Si need for rice. The budget of the dissolved silicon (DSi) fluxes gives the following results: the atmospheric and irrigation inputs represents 1% and roughly 10%, respectively, of the annual need for rice; the drainage and infiltration outputs represent 17 F 14 and 12 F 9 kg Si ha 1 year 1, respectively; the balance of our budget shows that at least 170 kg Si ha 1 year 1 are exported from the soil. If we consider the soil BSi as the only source of dissolved silicon, this stock could be exhausted in 5 years. D 2005 Elsevier B.V. All rights reserved. Keywords: Silicon cycle; Rice; Biogenic silica; Camargue 1. Introduction Rice and many other cultivated grasses are Si accumulators. As opposed to natural ecosystems, a substantial part of silicon accumulated in cultivated grasses does not return to soil. Silicon agricultural exportation is estimated to 210–224 million tons per year (Matichenkov and Bocharnikova, 2001), which is in the same order of magnitude as the total dissolved Si * Corresponding author. Tel.: +33 04 42 97 15 26; fax: +33 04 42 97 15 40. E-mail address: [email protected] (J.-D. Meunier). 0375-6742/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.gexplo.2005.08.036 transported by the rivers to the oceans per year. In order to determine the consequence of the decrease of available soil Si in cultivated areas, it is necessary to study the biogeochemical cycle of Si in grasslands and cultivated ecosystems as has already been done in forest ecosystems (Bartoli, 1981; Lucas et al., 1993; Alexandre et al., 1997; Markewitz and Richter, 1998; Farmer et al., in press), where plant Si dissolution was found to control the cycle of Si. Plant Si is mainly composed of phytoliths – amorphous, opal A particles – that are among the most readily weatherable forms of available Si in soils (Fraysse et al., 2004; Derry et al., 2005). Here we present a study of the role of biogenic silicon particles (BSi) and dissolved silicon (DSi) fluxes in the V. Desplanques et al. / Journal of Geochemical Exploration 88 (2006) 190–193 biogeochemical cycle of Si in a Camargue soil (SE of France) where rice has been cultivated for 21 years. 2. Material and methods 2.1. Study area The Camargue is the deltaic Rhône plain and is the main rice cultivation area in France. 105 420 tons of rice have been produced in 2003 with a mean yield of 5.6 tons ha 1 (FAO, 2004). The studied rice field is located in a large rice domain north west of the Vaccares lagoon. Rice was cultivated for 21 years between 1975 and 2003, whereas fennel and wheat were cultivated the rest of the time. Soils have developed on the fluvial deposits of an ancient Rhône channel. A compact plough sole is present between 25 cm and 40 cm (see Cary and Troland, 2006—this issue). Soils have a silty–sandy composition and the main minerals are quartz, calcite, K-feldspar and mica. The clay fraction contains illite, chlorite, smectite and mixed layer of smectite–illite. 2.2. Methods DSi input was analysed from rain waters collected in 2003 on the Plateau de l’Arbois (65 km east from our site). The volume of annual rain was calculated from the monthly average rainfall of the 1964 to 2003 period (Météo France at the Tour du Valat center). Irrigation waters were sampled during April 2005. DSi (b0.2 Am fraction) was analysed by UV spectrophotometry and ICP–AES. BSi (phytoliths, diatoms. . .) was extracted from 7 soil samples collected in a soil profile between 0 and 40 cm in depth using a wet extraction procedure (Kelly, 1990). Biomass and yield of the rice crop were measured at harvest by extrapolation from four 0.25 m2 plots. Roots were extracted from a 0.25 m2  0.25 m block of soil by humid sieving. Roots, straws and grains were dried at 105 8C and weighed. Silicon in plants (BSi) was determined by ICP–AES after melting with lithium metaborate. The particles were observed and analysed using an optical microscope and a scanning electron microscope (SEM) equipped with energy-dispersive spectrometry (EDS). precipitation, 6100 m3 ha 1 year 1, and 2) irrigation during rice cultivation (from April to September), 23 000 m3 ha 1 (Chauvelon, 1996); while water outputs are 1) infiltration, 6200 m3 ha 1 (from Godin, 1990; Chauvelon, 1996), and 2) drainage, 10 000 m3 ha 1 for five emergences per year. 3.2. Biogenic silica stock in soil The amount of BSi in soil ranges from 0.03 to 0.06 wt.%. These values are low compared to other soils (Clarke, 2003). Microscopic observations show that BSi particles are mostly phytoliths and diatoms. The surfaces of the phytoliths are covered by dissolution pits even in the top soil samples (Fig. 1). Many BSi particles, identified by EDS, have unidentified morphologies. The number of the unidentified BSi particles increases with depth. A similar trend has been previously observed and is due to a higher degree of dissolution of phytoliths at depth due to their longer residence time in soil (Alexandre et al., 1997). Calculation of BSi in the root zone (0–25 cm) above the compact plough sole, taking into account a soil density of 1.4 g cm 3 (Cary, 2005), gives 800 kg Si ha 1. 3.3. Global Si balance 3.3.1. Outputs – BSi in rice plants ranges from 0.9% (roots) to 2.5% (straw) DW. Taking into account the dry biomass of 17.4 tons ha 1, BSi removed by rice equals to 320 F 70 kg Si ha 1. We postulate that these values equal the Si rice needs to maintain the yield. The exportation of silicon from the field can be cor- 3. Results and discussion 3.1. Water mass balance Fluxes of DSi are calculated according to the water mass balance in the rice field. Water inputs are 1) 191 Fig. 1. Phytolith at 10 cm under the surface of the rice field. 192 V. Desplanques et al. / Journal of Geochemical Exploration 88 (2006) 190–193 Fig. 2. SEM photographies of irrigation waters residue. rected for agricultural practices. Harvest kept in field (roots and 10 cm of straws) represents 50 F 10 kg Si ha 1. Upper parts of the straws are heaped up and burnt. In Camargue, winds blow frequently and violently disperse ashes. Thus we have considered that the BSi contained in the upper parts of the straws left the rice field. Yet the amount of BSi returning to the soil is approximately 20% of the silicon taken up by the rice. Therefore, the growth of a rice crop exports 320 50 = 270 F 80 kg Si ha 1 year 1. – Assuming that DSi in infiltration waters equals 70 F 50 Amol L 1 (Moreau, 2004; Cary, 2005), the DSi output by infiltration is 12 F 9 kg Si ha 1 for a rice season. During the rest of the year, the rice plot is not flooded and we neglect infiltrations. – The average DSi concentration in flooded waters (60 F 50 Amol L 1; Moreau, 2004; Cary, 2005) combined with drainage output allows us to calculate a DSi output from drainage of 17 F 14 kg Si ha 1 year 1. – The dynamics of diatoms in the submersion waters is not well known. Batalla (1975) observed diatom development in Spanish rice fields but it was difficult to assess the diatom output and consequently the export of biogenic silicon. – The translocation in soil of biogenic silicon has not been evaluated yet. It should, however, be negligible because of the presence of a compact plough sole. 3.3.2. Inputs – In rain, the average DSi concentration is 15 Amol L 1. The annual precipitation flux gave a DSi input of 3 F 2 kg Si ha 1 year 1 which represented only 1% of the rice need. – In irrigation waters, DSi concentration is about 40 Amol L 1; this value gives a DSi input of 30 F 15 kg Fig. 3. Silicon transfers in a rice field of Camargue. V. Desplanques et al. / Journal of Geochemical Exploration 88 (2006) 190–193 Si ha 1 year 1 which roughly corresponds to 10% of the rice need. – SEM observations on the suspended matter of irrigation waters revealed the presence of diatoms (Fig. 2). The amount of diatoms transported by the Rhône River has not been assessed yet. As a rough approximation, diatom BSi was estimated from Garnier et al. (1995) and Davey (1986). The mean concentration of diatom BSi in irrigation waters during the rice cultivation period is estimated at 5 g Si m 3. The calculated diatom BSi input gives 100 kg Si ha 1, which corresponds to more than one-third of the rice need. Consequently, total (diatom BSi + DSi) input equals to 130 kg Si ha 1 year 1, whereas total (BSi + DSi) output is at least 300 F 100 kg Si ha 1 year 1 without the two unknown biogenic silicon outputs (Fig. 3). Thus, If we postulate that BSi in soil is the only source of dissolved silicon (dissolution of soil silicates is negligible), the stock of biogenic available Si will be exhausted in 5 years. 4. Conclusion We show that the input of Si in the rice field of Camargue (diatom BSi + DSi) accounts only for 44% of the rice needs. 56% must be provided by the soil constituents. If we postulate that BSi in soil is the only source of dissolved silicon (dissolution of soil silicates is negligible), the stock of biogenic available Si will be exhausted in 5 years. We have estimated that the irrigation waters bring more than one-third of the silicon need for rice as diatoms contained in irrigation waters. As silicon is essential to growth and rice development (Epstein, 2001), these preliminary results stress the need to assess more precisely the importance of BSi input as a natural fertilizer. Acknowledgements The authors wish to thank A. Alexandre, L. Bremond, Y. Noack, C. Pailles, F. Chalié, J. Bertrand, P. Moreau and X. Guillot for helpful discussions. This work has benefited from a support from the French program bECCO/PNBCQ. References Alexandre, A., Meunier, J.D., Colin, F., Koud, J.M., 1997. Plant impact on the biogeochemical cycle of silicon and related weathering processes. Geochimica Cosmochimica Acta 61, 677 – 682. 193 Bartoli F., 1981. Le cycle biogéochimique du silicium sur roche acide, application à deux écosystèmes forestiers tempérés (Vosges), PhD thesis, Université Nancy I. Batalla, J.A., 1975. Las Algas de Las Arrozales y El Empleo de Los Alguicidas. Publicado por la Federacion Sindical de Agricultores Arroceros de Espana (Internal report). Cary, L., 2005. Mobilité des éléments selon les alternances aéorobie anaérobie dans un écosystème rizicole en Camargue, PhD thesis, Université Paul Cézanne Aix - Marseille III. Cary, L., Trolard, F., 2006. Effects of irrigation on geochemical processes in a paddy soil and in ground waters in Camargue (France). Journal of Geochemical Exploration 88, 177 – 180. doi:10.1016/ j.gexplo.2005.08.033 (this issue). Chauvelon, P., 1996. Hydrologie quantitative d’une zone humide méditerranéenne aménagée: le bassin de Fumemorte en Grande Camargue, delta du Rhône , PhD thesis, Université Montpellier II. Clarke, J., 2003. The occurrence and significance of biogenic opal in the regolith. Earth-Science Reviews 60, 175 – 194. Davey, M.C., 1986. The relationship between size, density and sinking velocity through the life cycle of Melosira granulata (Bacillariophyta). Diatom Research 1 (1), 1 – 18. Derry, L.A., Kurtz, A.C., Ziegler, K., Chadwick, O.A., 2005. Biological control of terrestrial silica cycling and export fluxes to watersheds. Nature 433, 728 – 731. Epstein, E., 2001. Silicon in plants: facts vs concepts. In: Datnoff, L.E., et al., (Eds.), Silicon in Agriculture, Stud. Plant Sci. vol. 1, pp. 1 – 15. FAO, 2004. www.fao.org/ag/AGP/AGPC/doc/field/commrice/pages/ welcome.html. Farmer, V.C., Delbos, E., Miller, J.D., 2005. The role of phytolith formation and dissolution in controlling concentrations of silica in soil solutions streams. Geoderma 127, 71 – 79. Fraysse, P., Pokrovsky, O.S., Schott, J., Meunier, J.D., 2004. Surface properties, solubility and dissolution kinetics of phytoliths, from bamboos of Reunion Island. Special Supplement Abstract of the 13th Annual V.M. Goldsmidt Conference, Copenhagen, pp. 13. A216, 2.7. Garnier, J., Billen, G., Coste, M., 1995. Seasonal succession of diatoms and Chlorophyceae in the drainage network of the Seine River: observations and modeling. Limnology and Oceanography 40 (4), 750 – 765. Godin, L., 1990. Impact de l’irrigation pour la riziculture sur l’hydrologie et la chimie des eaux en Camargue. PhD thesis, Université Avignon et Pays de Vaucluse. Kelly, E., 1990. Methods for extracting opal phytoliths from soil and plant material. Documents of the Department of Agronomy, p. 10. Lucas, Y., Luizao, F.J., Chauvel, A., Rouiller, J., Nahon, D., 1993. The relation between biological activity of the rain forest and mineral composition of soils. Science 260, 521 – 523. Markewitz, D., Richter, D.D., 1998. The bio in aluminium and silicon geochemistry. Biogeochemistry 42, 235 – 252. Matichenkov, V.V., Bocharnikova, E.A., 2001. The relationship between silicon and soil physical and chemical properties. In: Datnoff, L.E., et al., (Eds.), Silicon in Agriculture, Stud. Plant Sci. vol. 13, pp. 209 – 219. Moreau, P., 2004. Biogéochimie des eaux en riziculture camarguaise, Master thesis, Ecole d’agriculture d’Angers.