Complex spring system was studied in Western Latvia using thermal imaging and orthophoto map. The study area is a protected geological monument containing group of springs with discharge 2-90 L/s. Freshwater limestone deposits are found there in the area of about 17 ha. The limestone was mined in 1930-ties and 1950-60-ties, but later abandoned pits were used for fish farming.
The aim of the study was to identify spring discharge and groundwater seepage locations in order to plan management of the area.
Handheld camera Flir E4 was used for thermal images. Drone DJI Phantom4 Professional with 2,54 cm 20MP camera sensor was used for aerial photography, from which orthophoto map was produced. Water temperature and conductivity were measured using WTW MultiLine P4 multimeter with TetraCon 325 electrode.
The study performed depends on season and air and water temperature. The field study was carried out in winter 2017-2018, when difference between air and groundwater temperature was at least 5 ℃ and there was no foliage or thick snow cover either.
The seepage of groundwater could be easily recognised on thermal images and their intensity depended on intensity of thermal signal. Thermal mapping of springs is challenging in underwater groundwater discharge places, where surface water temperature masks springs. The ratio of pond size and the spring discharge intensity matters in this case. The simultaneous water temperature measurements showed that in case of groundwater seepage and intensive spring discharge the thermal camera temperature measurements and manual measurements were close, within 0.5 ℃ range, but in case of underwater discharge the difference could reach 1-2 ℃. Therefore, thermal camera measurements cannot replace manual temperature measurements when precise data is needed, however, thermal mapping helps to pinpoint places where direct measurements should be conducted.
The season represented in orthophoto with captured floral characteristics of current spring ecosystem expressed as active plant growth near springs gives the color signature suitable for image classification and automatic delineation of these zones. Two approaches were tested and compared - unsupervised image classification and supervised classification using manually classified training polygons. Both approaches helped identify groundwater discharge locations, and worked well in underwater discharge cases as well.
This study showed that use of handheld thermal imaging camera provides good results in terms of identification of groundwater seepage zones and is rather effective in case of underwater springs. Combining thermal images with orthophoto analysis helps pinpoint places for detailed investigations in poorly accessible area.