Estimation of lag time of water and nitrate flow through the vadose zone: Hauraki catchment

Abstract

This report predicts the travel time for nitrate to travel from the land surface, though the unsaturated (vadose) zone and into shallow groundwater. The area of study is the entire Hauraki catchment and also the Coromandel peninsula, an area of 588,742 hectares. The travel time through the vadose zone can be an important component of the overall lag time between land use changes and associated surface water quality impacts. The modelling of vadose zone lag times in this report consists of four main stages of calculation: 1. a simple canopy interception model to estimate effective rainfall and PET 2. a soil moisture balance model to estimate land surface recharge 3. an unsaturated flow model to estimate vadose zone travel time 4. a saturated flow model to estimate the time taken for the water and nitrate to penetrate into the uppermost aquifer layer In this report the results of the vadose zone and saturated zone models are collectively referred to as the total travel time. The input data for these calculations has been sourced from available climate, soil, geological, and hydrological databases. The modelling results predict vadose zone median transit times of 10.9 years, and total transit times (allowing for groundwater mixing) of 14.6 years. The results are largely influenced by the depth to the water table, so most of the Hauraki Plain has total predicted travel times less than 10 years, and approximately half of the land below 50m elevation has transit times of 1 to 2 years. Much of the land at high elevations along the Coromandel and Kaimai ranges has deep predicted groundwater levels, and total travels times in excess of 30 years. The lack of available groundwater level data in these areas means the travel time predictions in these areas need to be treated with caution. The saturated (mixing) component of the calculations comprises around 26% of the total travel time. The calculation of the mixing time enables the model results to be compared with mean residence times derived from tritium samples. The model results are comparable to the tritium data, which provides some confidence for transit time predictions less than 25 years. A hypothetical irrigation scenario was carried out at all sites with land use classed as high producing pasture, orchard and crop. The irrigated simulation increased annual recharge by 15%. The respective transit times for these sites decreased by 19% on average under irrigation. The simulation showed that irrigation not only reduces the travel time, but also reduces the variability resulting from rainfall recharge.