4.1 Impact of subsurface drainage on the water balance
From Sept. 2019 to Aug. 2020, surface and subsurface flows represented 11.3% and 8.8% of the total annual rainfall, respectively. The simultaneous measurements of subsurface and surface drainage discharges during an entire year in a field drained by temporary artificial rill and tile drains is an originality of the present study. Previous studies quantifying both surface and subsurface drainage discharges have been conducted on fields only drained by tile drains (Schwab, G. O. et al., 1980; Skaggs et al., 1994; E. Turtola & Paajanen, 1995; Eila Turtola et al., 2007). In these studies, surface runoff pathways are free. In the present study, surface runoff is oriented manmade SDRs. So the surface runoff measurement is realized for a precise area that is the same all along the study period. Nevertheless, present results suggest that the surface runoff guidance by artificial rills do not modify the relative contribution of the two drainage networks: some studies report that the subsurface runoff represents 34% to 68% of the total runoff.
In order to have a better understanding of the hydric functioning of the field, a water balance could be established for the year 2019-2020. The calculation of the water balance (Figure 8) was performed on a daily time step with the components presented in Equation 2: precipitation (P), real evapotranspiration (RET), surface and subsurface runoff (D), and soil infiltration. Soil infiltration is composed of two terms: the variation of the water stock of the LA and S1 surface horizons (ΔS) and the infiltration to the underlying horizon (I), from S1 to S2. As there is no measurement of the variation of the water stock of the LA and S1 horizons throughout the year, the variations were estimated. The hydric capacities of the horizons were estimated from a pedotransfer function proposed by Saxton (1986) and based on grain size. The horizons (LA+S1) have an estimated cumulative total available water of 75.68 mm. The estimation of water transfers to LA and S1 is based on the following assumptions:
- (H1) The water stock can not exceed the water stock at field capacity - (H2) The water stock can not be lower than the water stock at the permanent wilting point
In addition, the distribution of water inputs or losses is based on the following hypothetical processes:
- (H3) When the effective rainfall is positive, water input is first allocated to the LA and S1 horizons provided that it does not contradict assumption (H1), then to the S2 horizon.
- (H4) When the RET is higher than the rainfall, the water losses are allocated in priority to the LA and S1 horizons, provided that they do not contradict hypothesis (H2), and then to the S2 horizon.
- (H5) At the beginning of the hydrological year, the soil water content is considered to be close to or equal to the water content at the permanent wilting point, both for the (LA+S1) horizons and for the underlying horizon.
The discharges and stock changes calculated under these assumptions are presented in Figure 8. After the rains of September and October, the water content of the LA+S1 horizons reaches the water content of the field capacity. This is confirmed by manual tensiometer measurements taken in early November. Variations in the measured depth of the saturated zone even indicate that the LA and S1 horizons can be rapidly saturated during a rainfall event, which is consistent with a pre-event water content close to field capacity. Once the water content of the LA+S1 horizons reaches the field capacity water content, flows from both surface runoff and tile drains were monitored during some rainfall events. Flow from tile drains becomes continuous from late November to mid-December (Figure 5.). At the same time, water percolation from S1 to S2 begins.
This is consistent with manual measurements of soil tension: at 95 cm depth, soil water tension is divided by 4 between early November and mid-December. By mid-December, 62% of the volume of water infiltrated to S2 during the intense drainage period has percolated. Until this date, percolation from S1 to S2 can represent up to 96% of the daily precipitation. Little percolation was calculated in January, suggesting that the S2 horizon was saturated. At the end of the intense drainage period, the volume of water percolated from (LA+S1) to S2 is 1.4 times the volume of water drained: 193.8 mm and 139.5 mm, respectively. By mid-March, no tile drain flow was measured, because of decreased soil water content in the LA+S1 horizons. This result was confirmed by experimental measurements. From mid-March, the water stock in the LA and S1 horizons did not reach the water stock at field capacity and no transfer from S1 to S2 is calculated. S2 water stock progressively decreases from mid-March to the end of August. Over the entire hydrological year, S2 water stock increased by 50.7 mm and subsurface drainage captures 54% of the water transiting through the soil (Table 5).
The water balance highlights the importance of percolation from S1 to S2 in the distribution of flows within the field. In particular, it seems that the clayey S2 horizon plays a buffering role in the initiation of subsurface runoff: subsurface runoff intensifies when the storage capacity of the S2 horizon is reduced. However, almost the half of the flow is not intercepted by the subsurface drains.
All of the assumptions made for the calculation of the water balance appear to be consistent with the measurements made during the intense drainage period. However, these assumptions, and more particularly assumptions (H3) and (H4), are less suitable for the low drainage period. Indeed, the development of the crop and the intensity of evapotranspiration mean that water loss must certainly affect the LA/S1 and S2 horizons in a more distributed manner. In addition, the presence of cracks significantly modifies the distribution of water in the soil (Koivusalo et al., 1999) and may lead to direct flow to S2. The seasonal contrast of the soil water content in the superficial horizons suggest changes in soil structure that may have consequences on the soil water pathways. In the following part, soil water pathways under dry and wet conditions are described.