NEUTRON PROBE MEASUREMENT OF SOIL WATER CONTENT CLOSE TO SOIL SURFACE

The problem of neutron probe soil water content measurements close to soil surface is analysed from the spatial variability and also from the slow neutron "loss" to the atmosphere points of view. Results obtained on a dark red latosol of the county of Piracicaba,SP, indicate the possibility of precisely measuring the neutron "sphere of influence" when different media are used on soil surface.


INTRODUCTION
The principle of soil water content measurement with neutron probes involves the establishment of a slow neutron "cloud" in the soil around the probe, which defines its "sphere of influence".For measurements close to soil surface, part of the sphere is in the atmosphere, causing a loss of slow neutrons and consequently incorrect measurements.It is therefore of great importance to know the radius of the sphere to define how close to soil surface measurements can be taken without neutron loss.Several authors, e.g.HAVERCAMP et al. (1984), MENDES et al. (1992) and TURATTI & REICHARDT (1991) avoid surface measurements and assume surface water content as being equal to the first measurement made in depth.KIRDA & REICHARDT (1992) compare neutron probe measurements with other non-nuclear methodologies, but avoid discussing surface measurements with depth probes.GREACEN (1981) describes extensively the assessment of soil water content by the neutron method, and for the case of surface measurements, suggests the elaboration of special calibration curves or the use of specially designed surface equipment.This study contributes to the understanding of neutron probe measurements close to soil surface, suggesting a new approach to define how close to soil surface safe water content measurements can be made and also quantifying the spatial variability of these measurements, for a given soil type.
3. EXPERIMENTAL DESIGN: two plots of 7 x 7 m, two meters apart, each with 9 aluminum access tubes, as shown in Figure 1.These access tubes were used to make neutron probe readings from above soil surface down to the 50 cm depth.Measurements were performed with the upper part of the access tube in air, in plastic (using a factory designed "deflector") and in water (using a pool avoiding water infiltration into the soil).Measurements were performed in two soil water content conditions: i) dry condition, in the dry season, when water content reaches miminum values in the field, and ii) wet condition, in the rainy season, when water content reaches maximum field values.

RESULTS AND DISCUSSION
Figures 2 and 3 present soil water content data obtained from the access tubes having the upper part in air, as recommended.The figures give an idea of the variability of soil water content measurements which is quantified in TABLE 1.The tendency of an increase in C.V. from the 50 cm depth in direction to soil surface might be due to the erratic loss of neutrons to the atmosphere.It is difficult, however, to separate this effect from water content and bulk density effects, since these parameters are more variable close to soil surface.
From the depths greater than 30 cm, for which one can be sure that the slow neutron cloud is completely inside the soil, coefficients of variation are of the order of 4.5% for the dry condition and of 3.0% for the wet condition.TABLE 1 also presents the number N of measurements necessary to obtain a new mean value within d% deviation of the estimated true mean.N was calculated according to WARRICK & NIELSEN (1980) using the equation: where x a = 1,96 and a 2 is the variance.The N values shown in TABLE 1 are a guide for the most common question which is the recommendation of number of access tubes needed.Data shown in Figure 4 are count-ratios (CR), obtained dividing slow neutron counts in soil, soil-air and air, by counts in water.As the probe moves from soil to air, slow neutrons are lost and CR values tend to zero.The problem lies in finding the depth below which one is sure that there is no neutron loss, thus denning the sphere of influence of the probe.If the soil is homogeneous and at constant water content in depth, CR values are constant in depth if the sphere of influence is completely inside the soil.Small deviations from this value indicate that part of the sphere reached the atmosphere, and its radius can be estimated graphically, as shown in BACCHI & REICHARDT (1990).
In most cases, however, soil is not homogeneous and mostly in terms of water content.It is therefore not possible, in situations like Figure 4, to define the radius of the sphere.On the other hand, if a medium of high neutron moderation characteristics is placed around the access tube, above soil surface, the definition of the influence sphere becomes very evident.Figure 5 is an example using a plastic factory made neutron "deflector" or a water pool around access tube at soil surface.When part of the slow neutron sphere reaches this different medium of high moderation property, count ratios at soil surface increase rapidly in relation to measurements taken in the air.
Although the effect becomes more evident for the water pool, this procedure is very unpractical to be performed under field conditions for a large number of access tubes.The factory made deflector or any plastic or paraffin block are very handy and can be used successfully to define the shallowest depth at which safe measurements can be made using neutron probes.
In the case of Figure 5 it can clearly be seen that the radius is of the order of 20.0 cm for the soil.Measurements were also taken with and access tube in water (Figure 5) showing that in this case the radius of the sphere of influence is of the order of 10.0 cm.