Fibre Optic Distributed Temperature Sensing (FODTS) to Characterize Groundwater/StreamInteraction in New Zealand Hydrogeological Settings

Abstract

SMART (Save Money And Reduce Time) aquifer characterization (SAC) research programme was established to identify, develop, apply, validate and optimise a suite of highly innovative methods for accurate, rapid and cost-effective characterization and mapping of New Zealand's groundwater systems. The methods include: ambient noise seismic tomography; airborne geophysical surveying; satellite remote sensing; fibre optic distributed temperature sensing (FODTS); and novel age tracers. Validation of each techniques is achieved by the use of multiple methods in one case study and=or by ground-truthing to existing data collected from traditional methods. Implementation of the techniques are followed by quantification of uncertainty and data visualisation. This study is one of four PhD projects in the SAC research programme and for the first time in New Zealand, investigates the applicability of FODTS technique in a range of rivers and streams to characterize groundwater/surface water interaction. In this project, four case studies were undertaken including the Ngongotaha Stream, the Blue Spring section of the Waihou River, the Tutaekuri-Waimate Stream and the Kahahakuri Stream. The selected streams/rivers were located in typical New Zealand hydrogeological settings including volcanic and alluvial aquifers and cover a range of base flow regimes from ~100 to 2000 L/s with different morphologies and gaining systems such as discharge through the streambed, discharge from the stream banks, diffuse discharge and spring-fed tributaries. The main objectives of using FODTS in the Ngongotaha Stream was to identify the location of known springs as a means of testing the technique. Moreover, to see if the technique with high spatial temperature measurement capability could improve existing understanding of groundwater/surface water interaction in the stream by finding new springs or gaining reaches and eventually use the collected temperature data to quantify groundwater discharge. Fibre optic cable was deployed near the true left and right banks of the Ngongotaha Stream over a 973 m reach with the average flow of ~800 L/s (in the study reach). Groundwater discharge locations in the study reach were identified by FODTS profiling (using the constant temperature and the standard deviation of diurnal temperature methods) and visual reconnaissance. Thirteen springs/tributaries were detected, five discharged from the true right bank and eight from the true left bank. Previous studies in the Ngongotaha Stream identified ten springs in the study reach (Kovacova et al., 2008). To quantify groundwater discharge in the Ngongotaha Stream, a new approach was developed in this study, in which the one dimensional transient heat transport model was fitted to the FODTS measurements, where the main calibration parameters of interest were the unknown spring discharges. The spatial disposition of the groundwater discharge estimation problem was constrained by two sources of information in this study; firstly, the stream gains ~500 L/s as determined by streamflow gauging. This provides a total volume of groundwater discharge in the study reach. Secondly, the temperature profiles of the left and right banks provide the spatial disposition of springs and their relative discharges. In this way the spring discharges were quantified in the relatively complex setting of the Ngongotaha Stream. Fibre optic cable was deployed near the centre line of the Waihou River, the Tutaekuri- Waimate Stream and the Kahahakuri Stream over 1235, 1592 and 2100 m reaches, respectively. To identify springs or gaining reaches and to quantify groundwater discharge, the constant temperature method and the steady-state heat transport model were used, respectively. In the Blue Spring section of theWaihau River the aim was to test of the suitability of the technique in a relatively large deep river ( ~ average flow of 2000 L/s) with one large known spring ( ~500 L/s discharge) flowing from a fractured volcanic aquifer. The FODTS data analysis could identify the location of the Blue Spring. It is possible that more springs occur in the study reach that were not identified on the FODTS profiles. The Blue Spring discharge was estimated at 560 L/s in this study. The previous synoptic gaugings in upstream and downstream of the Blue Spring estimated the discharge at 498 L/s (Van Kampen, 2001) and 680 L/s (Hadfield, 2011). The spring discharge that quantified using the steady-state heat transport equation in this study was within the range of these two gauged estimations. Prior to the deployment of FODTS equipment in the Tutaekuri-Waimate Stream, it was known from concurrent gauging (Dravid et al., 1997), that streamflow increased downstream due to groundwater inflow but the location, spatial distribution of flux and the mechanism of gain were unknown. Visual inspection of the stream undertaken by this study, did not identify any discrete groundwater sources except a small groundwater-fed tributary. Fibre optic cable was deployed in the Tutaekuri-Waimate Stream with the average flow of ~1000 L/s (in the study reach) located on a boundary of a gravel and limestone groundwater systems. Two diffuse groundwater inflow sections and two springs were identified fromthe FODTS temperature profiles. A tributary which was observed during field investigation was also identifiable on the FODTS profile. The amount of groundwater inflow in two diffuse sections and two discrete springs was estimated at 357 and 173 L/s, respectively. The total gain in streamflow in the study reach, including the tributary inflow, was estimated at 613 L/s using the steady-state heat transport model. The streamflow gaugings indicated a 590 L/s increase in flow over the study reach. Prior to the deployment of FODTS equipment in the Kahahakuri Stream, it was known that part of the Waipawa River water loss to groundwater is captured by the Kahahakuri Stream (Undereiner et al., 2009). However, the gaining system, location and the distribution of groundwater inflow into the stream had not been studied. FODTS was deployed in the Kahahakuri Stream with the average flow of ~100 L/s in the study reach located on a gravel aquifer system. In the study section of the Kahahakuri Stream, only one spring was found from the FODTS temperature profiles. It appears that the Kahahakuri Stream gains most of its flow through spring-fed tributaries not by direct groundwater inflow to the stream. Due to the close proximity of tributaries 2 and 3 confluences with the Kahahakuri Stream, and their temperature differences (tributary 2 was generally warmer (21- 24°C) than tributary 3 (14-17°C)), the steady-state heat transport equation could not be used to estimate the discharge of these tributaries. The estimated discharge for the spring and tributary 1 using the steady-state heat transport equation were 3.4 and 167.5 L/s, respectively. The total discharge of the spring and tributary 1 (170.9 L/s) was consistent with the synoptic gauging result (170 L/s). FODTS technique has shown very good potential in the case study streams/rivers by identifying groundwater discharge locations through distributed temperature measurements at very high spatial resolution. The temperature profiles measured by FODTS were used successfully in the steady-state and transient heat transport models to quantify groundwater discharge. Application of FODTS technique, improved the existing understanding of groundwater/surface water interaction by providing detailed characterization of the groundwater inflow in the range of hydrogeological setting of the study areas.