Unlike rice season, rainfall is the only driver of water leaching
in wheat season. Water leaching occurred randomly because
the rainfall amount, intensity and frequency varied greatly. This
made the observation of amount of leaching water so difficult in
the field scale. So we used field undisturbed tension-free monolith
lysimeters (Fig. 3c) to directly obtain the total amount of leaching
water in wheat season. Three lysimeters were installed and cultivated
exactly like the rice or wheat plants in the field experiment
under the same water and fertilization management. A simulated
underground water supply system could provide a 10 cm height
water layer in the bottom of lysimeters and prevent drought cracking
of the soil monolith caused by evaporation and transpiration.
Two runoff outlets were also installed at 0 and 15 cm soil depths
to simulate runoff on the soil surface and from the 15 cm deep
drainage ditches. These special practices ensured that the amounts
of leaching water obtained from the lysimeters were more realistically
similar to those in the field. Although certain errors remain,
the data obtained by lysimeters are at least measured values from
the field, varying from estimated values using the water balance
method.