Scientific registration n : 1368 Symposium n : 3 Presentation : poster Rainfall infiltration into a volcanic ash soil and soil water flux at 1-m depth Infiltration de la pluie dans un sol cendreux volcanique et flux d eau à une profondeur de un mètre HASEGAWA Shuichi National Institute of Agro-Environmental Sciences, Tsukuba, 305 Japan INTRODUCTION Many studies conducted to characterize water movement in field soils have indicated that preferential flow often occurs. Such preferential flow is divided into two types. Bypass flow occurs often in structured soils in which water flows rapidly through cracks, fissures and biopores. Funnel flow or non-uniform infiltration is affected by the surface microrelief and vegetation. Time Domain Reflectometry ( TDR ) allows the monitoring of soil-water content in situ without destruction of the soil profile. If the increase in soil-water content in a soil profile is equivalent to the amount of rainfall, water flows uniformly into the soil matrix but, if the amount of rainfall exceeds the increase of the amount of soil water storage, surface runoff and/or preferential flow occurs. Preferential flow leaches dissolved solutes to a greater depth than occurring by uniform infiltration into the soil matrix. As a result, the opportunity for plants to absorb nutrients decreases, resulting in possible contamination of groundwater. Therefore, estimation of the uniform infiltration and preferential flow rates is very important for environmentally conservative agriculture. Volcanic-ash soils, which occur in approximately half of the arable upland in Japan, are characterized by a low bulk density, high water content at -0.1 and -1.5 MPa, high saturated ( >10-3 cm s -1 ) and unsaturated ( >10-6 cm s -1 at -10 kpa ) hydraulic conductivities ( Hasegawa et al., 1994 ) and abundant biopores formed by decayed plant roots in the subsoil. The purposes of a present study are 1: to monitor soil water content to a depth of 1 m in an unvegetated volcanic-ash soil throughout a year and to determine the effects of surface microrelief on non-uniform infiltration and the amount of matrix flow in relation to the annual precipitation, and 2: to measure soil water flux at 1-m depth to evaluate amount and displacement of water which carries inert and nonadsorbing solutes. 1
EXPERIMENTAL PROCEDURES Experimental site Experiments were conducted in a field at the National Institute of Agro- Environmental Sciences ( latitude 36 degrees and 1 minute N and longitude 140 degrees and 7 minutes E ). Annual mean precipitation and temperature for 1981 to 1990 were 1219 mm and 13.0 degrees centigrade, respectively. Daily minimum temperatures during December to February were below 0 degrees centigrade so that the surface soil was frozen in the morning but thawed in the afternoon. The soil to a depth of 200 cm is a volcanic ash, classified as a Hydric Hapludand ( Soil Survey Staff, 1992 ), with a texture of heavy clay. The soil profile is divided into plough layer ( 0-20 cm ), ploughsole ( 20-30 cm ) and subsoil ( >30 cm ). No cracks and fissures were visible at the surface, but a large number of tubular pores formed by decayed plant roots were found in the subsoil. Field measurements Measurements of the soil water content by TDR and of rainfall by a tipping bucket rain gauge were started on 1 December 1994 and terminated on 13 December 1995 at intervals of every 30 or 60 min. The TDR system was a Tektronix 1502B cable tester interfaced to a Campbell Scientific 21X data logger with PROMS software. The TDR probe consisted of two parallel stainless rods 5 mm in diameter and 50 mm apart. Four 1-m probes were installed vertically at intervals of 2 m and three 30-cm probes were installed between the 1-m probes. The experimental site was kept weed free throughout the year by hand weeding. This TDR measuring system was used again in 1997. Three 1-m and 30-cm vertical probes from the soil surface and two 30-cm horizontal probes at 1-m depth were installed in a maize-chinese cabbage cropping field. Soil water suctions at depths of 90 and 110 cm by triplicate tensiometers and TDR were measured every 30 min.. Soil core samples were collected at four different times to convert the TDR water content to oven dry water content. Thin-walled steel tubes of 110 cm long were driven into the soil to a depth of 1 m and retrieved. Laboratory experiments Three undisturbed soil core samples, 10.0 cm in diameter and 4.0 cm high, were taken at a depth of 100 cm. Unsaturated hydraulic conductivities both for drying and wetting processes were measured in a range between a few cm and 150 cm suctions by a steady state method after Richards and Moore ( 1952 ). Soil water characteristic curve by a suction plate method for drying and wetting processes were also obtained for the same soil cores used for unsaturated hydraulic conductivities. RESULTS AND DISCUSSION Determination of soil water content by oven dry and TDR methods Volumetric soil water content of layered soil by the oven dry method was calculated by the product of gravimetric water content and dry bulk density of the each layer. For the TDR water content, we used an empirical equation between the relative dielectric constant of soil and the volumetric water content ( Topp et al., 1980 ). Soil water contents determined by the core sampling method were always higher than those calculated using the Topp s equation. However, the coefficients of determination of the 2
linear regression lines between core sampling and empirical methods were high enough for all the probe lengths. Annual changes of soil water content Annual precipitation from 1 December 1994 to 30 November 1995 was 1192 mm, close to the annual average of 1219 mm. Soil water content to a depth of 1 m during the winter season from December 1994 to February 1995 was low and relatively constant due to small amount of precipitation and soil evaporation. Monthly rainfall from March to July exceeded 100 mm and soil water content in this period was high. After the end of the rainy season in the fourth week of July, the soil began to dry and the driest condition appeared immediately before a typhoon on September 15, and then it became the wettest by the typhoon. Field capacity defined as water stored to 1-m depth on the next day after a heavy rain was 640 mm, and the difference of water between the field capacity and the driest condition was about 100 mm. Increase of amount of storage water by rainfall Infiltration characteristics of rainfall were investigated for 36 rainfall events exceeding 10 mm with a total amount of rain of 969 mm for an annual rainfall of 1192 mm. In case of 25 rain events with the intensity less than 10 mm h -1 into dry soil, rainwater infiltrated uniformly into the soil matrix increasing the soil water content. The increase in the amount of storage water to a depth of 1 m enabled to predict the cumulative rainfall with standard deviation of 1.7 to 4.3 mm. In 6 rain events, the intensity was higher than 10 mm h -1 and surface runoff to the depressions occurred resulting in their excess water infiltration vis-à-vis the amount of rainfall. Lower parts received 1.4 to 1.6 times more rainfall than the higher parts, even though surface microrelief among the probes was in the range of 6 cm ground height. Furthermore, most of rainfall was absorbed and stored in the surface 30-cm layer, and bypass flow might not have occurred in the subsoil. There were two large rainfall events in 1995. In one, 65.4 mm of rain fell on from 15 and 16 May following 40.8 mm on 12 and 13 May. In the other, 155.2 mm fell as a result of a typhoon on 16 and 17 September. In these rains the difference between the amount of rainfall and the amount of storage water were 25 and 40 mm, respectively. Therefore, the total of 65 mm of water did not increase the soil water content. Even though we did not determine whether the field was completely flooded or not, surface runoff might have carried away a certain amount of rain. Under the condition near saturation, bypass flow is likely to occur through continuous macropores whose volume is very small compared to the volume in the soil matrix. The amount of bypass flow was, definitely, less than 65 mm or 5.5% of the annual rainfall when the surface runoff was subtracted. This must be one of a distinct characteristics of the volcanic ash soil unlike other fine textured soils. Little shrinkage property by drying and lack of strong dry spells impede development of large drying cracks. Soil matrix having high saturated and unsaturated hydraulic conductivity conducts infiltrating water rapidly and has chance to absorb water flowing through the macropores into the soil matrix. These must be major reasons that bypass flow seldom occurred for the volcanic ash soil. Soil water flux 3
Suction and hydraulic conductivity relation, ( p-k ) and soil water characteristic curves, ( p-θ ) had a hysteresis between drying and wetting processes. However, a relation between k and θ obtained from p-k and p-θ showed substantially no hysteresis. This enabled to convert soil water content by TDR to hydraulic conductivity directly. Soil water flux at 100-cm depth was calculated by Darcy s law, in which hydraulic gradient was obtained by tensiometers installed at 90- and 110-cm depths. Water balance Cumulative soil water flux by Darcy s law was examined by a water balance method. R = S + E + Q (1) where R is amount of rainfall, S is change of stored water to a depth of 1m, E is evapotranspiration and Q is the cumulative soil water flux for a period considered. In applying the water balance equation, following conditions were taken into account. (a) As TDR is not able to detect changes of stored water to a depth of 1 m of the order of 1 mm, the period that Eq. (1) is applied should be long enough. (b) Intensity of rainfall during a given period should be low enough to secure uniform infiltration into the soil matrix. (c) Evapotranspiration is the only factor that we could not measure under field conditions and, therefore, should not be a dominant term in the water balance equation. Considering above conditions, Eq. (1) was applied for 29 November to 13 December in 1997. Total amount of rainfall during this period was 53.3 mm and maximum intensity of rainfall was 6.5 mm h -1. Stored soil water to a depth of 1 m increased 6 mm on 13 December compared to 29 November and cumulative downward soil water flux was 35.5 mm. Amount of evapotranspiration calculated by Eq. (1) was 14 mm for 15 days or 0.93 mm d -1, which value must be reasonable considering that soil surface was frozen in night time. This result proved that our method to determine soil water flux was appropriate. Changes of soil water flux across 1-m depth Annual precipitation in 1997 was 974 mm which was lower than the long term average of about 1200 mm. Appreciable downward flux occurred after heavy rain events or rains concentrated in a short term. Soil water flux was always downward in June but was always upward in August. Upward soil water flux occurred about a half of days in the year, but total soil water flux was 161 mm downward. At two heavy rain events on May and June, cumulative soil water fluxes by Darcy s law were lower than those from water balance equation. This suggests that surface runoff and/or preferential flow occurred. As the increase of stored water by TDR under high intensity rainfall was not always equal to the amount of rainfall as mentioned previously, it is difficult to apply the water balance equation and to estimate separately the amounts of the surface runoff and the preferential flow. We assumed that the surface runoff as a whole did not occur and that difference of stored water between the maximum and at 24 hours after a rain subtracted by downward flux ( matrix flow ) was the preferential flow. The preferential flow was thus estimated to be 56 mm for the two rain events. Therefore, total amount of downward flow was calculated to be 217 mm. Total amount of downward flux divided by the average soil water content at 1- m depth shows a displacement of inert solute for one year. As average soil water 4
content throughout a year was 0.665 cm 3 cm -3, solute existed at 1-m depth moved down to 124-cm depth for one year. However, effect of preferential flow ( 56 mm ) on the displacement of solute was difficult to be evaluated. Amount of evapotranspiration in our region has not been studied precisely but was estimated between 700 and 800 mm per year using a water balance method to some watershed areas. Annual evapotranspiration of 754 mm in 1997 by Eq. (1) might be an appropriate value. CONCLUSION Soil water conditions and movement were studied using the TDR method in a field consisting of volcanic ash soil. Stored water was well monitored by TDR, which enabled to analyze the soil water conditions such as field capacity, stored water in the root zone and intensity and duration of dry spells. Low intensity rainfall to dry soil increased the amount of storage water in parallel to the cumulative rainfall measured with a rain gauge. When the intensity of rainfall was higher than 10 mm h -1, nonuniform infiltration influenced by the surface micro-relief occurred. A large amount of rainfall exceeding 100 mm including previous rainfall must cause surface runoff and /or bypass flow but contribution of bypass flow was estimated to be less than 5.5% of the annual rainfall in 1995. There was little hysteresis between hydraulic conductivity and soil water content relation. Matrix flow was measured by tensiometers and hydraulic conductivity converted from TDR soil water content in the field. Cumulative soil water flux by this method after rainfall coincided well with that obtained from a water balance equation. The amount of the matrix flow was 161 mm downward and 56 mm of the preferential flow added to it in 1997. Annual rainfall was 974 mm and evapotranspiration was estimated to be 754 mm. References Hasegawa, S, Osozawa, S. and Ueno, H., 1994. Measurement of soil water flux in Andisols at a depth below a root zone of about 1 meter. Soil Sci. Plant Nutr., 40: 137-147. Richards, L.A. and Moore, D.C., 1952. Influence of capillary conductivity and depth of wetting on moisture retention in soil. Trans. Am. Geophys. Union, 33:531-539. Soil Survey Staff, 1992. Keys to soil taxonomy, 5th ed., p.159, Pocahontas Press, Inc., Blacksburg, Virginia. Topp, G.C., Davis, J.L. and Annan, A.P., 1980. Electromagnetic determination of soil water content: Measurement in coaxial transmission lines. Water Resour. Res., 16: 574-582. Keywords: infiltration, matrix flow, TDR, water balance, in-situ, measurments, Japan Mots clés : infiltration, flux matriciel, TDR, bilan hydrique, expérimentation in-situ, Japon 5