Abstract
Soil as a three-phase system (solid particles, water and air) carries vital functions in the global ecosystems and is crucial for human societies as it acts as a medium for plant growth, purifies water for consumption by controlling the transport of contaminants, affects the earth’s atmosphere and is a habitat for many living organisms. For agricultural production, all three phases need to co-exist, hence the need for drainage of agricultural lands has long been realized, studied and practised by different civilizations over history. The widespread adoption of subsurface drainage practices started in the mid-19th century and since then they flourished concurrently with the intensification of agricultural activities to meet the growing population demand. Today, some of the most productive regions across the globe are a result of these artificial drainage systems that are mainly installed to remove the excess water in humid regions and prevent soil salinization in arid and semi-arid regions.
Agricultural subsurface drainage systems provide many agronomic, economic, and environmental benefits. However, if improperly managed they have deleterious effects on the environment as they shorten pathways for the transport of solutes to the aquatic ecosystems. Thus, knowledge of these buried drainage pipes’ location is crucial to comprehend the hydrology and solute dynamics and plan effective mitigation strategies. This knowledge can also benefit drainage water management, a conservation practice that is increasingly being adopted to modulate the water table where a control structure can be retrofitted to the existing drainage system and on the contrary, for the drainage intensification, as it is a typical practice to install the new drainage system in between the existing one. Despite this, the location of the drainage systems is often poorly documented or entirely unknown.
Conventional methods (tile probes and trenching equipment) for finding drainage pipes are invasive, tedious and causes pipe damage. Therefore, there is a growing need to find rapid, effective and non-destructive methods for drainage mapping purpose. Previous research showcases that proximal and remote soil and crop sensing methods may provide a potential alternative solution – amongst which maximum success was achieved by employing a ground penetrating radar (GPR) and aerial imagery captured using different cameras. Recent technological advances led to rapid developments in these techniques. While this constitutes the development of a 3D-GPR capable of scanning an ultra-wideband frequency with an antenna array of wide-area swathe, on another front, the unmanned aerial vehicle (UAV) and their associated cameras have become cheaper than ever making them flexible and feasible to capture aerial imagery. Hence, this dissertation aims at mapping subsurface drainage systems with state-of-the-art proximal and remote sensing methods (i.e. GPR and UAV imagery). A novel magnetic gradiometer (tMag) was also used but only to a limited extent and further research was discontinued owing to the lack of success.
In relation to the stepped-frequency continuous wave 3D-GPR, firstly, at one study site in Denmark, the performance of the system was assessed for different survey configurations (i.e. ground-coupled vs air-coupled) and site conditions (i.e. dry vs wet soil) to discern optimal setup and timing for performing the drainage mapping surveys (Paper V). Out of the combinations tested, a clearer drainage pipe response (i.e. hyperbolic patterns in the vertical profiles) was observed in the survey carried on wet soil with snow cover, however, the ringing noise was substantial at places and prevented the detection of the shallow drainage pipes. The surveys performed on dry soil showed a similar drainage pipe response in both the survey configurations. Later, a thorough assessment of a 3D-GPR was made based on a study conducted at 12 sites in Denmark on a variety of soil types (Paper I). Here, algorithms were developed for calculating global and localized penetration depth (PD). Two different approaches were tested for determining the localized PD. While the first approach was qualitative and helpful to comprehend the attenuation characteristics of the subsurface, the second approach permitted quantification and allowed comparison between different study sites. In both approaches, efforts were made to comprehend the support extended by the electrical conductivity (EC) measured using electromagnetic induction (EMI) instrument to evaluate the performance of the 3D-GPR in finding the drainage pipes. Overall, a high success rate was observed at five amongst the 12 sites visited. At seven sites, the 3D-GPR demonstrated less success owing to a high soil EC (thus limiting the 3D-GPR PD) and the driving direction being parallel to the drainage pipe orientation.
In relation to the UAV imagery, the studies conducted in Midwest U.S.A showcases the swiftly evolving UAV technology and associated cameras as a cost-effective and feasible tool for the drainage mapping purpose. More specifically, to determine the overall feasibility, a comprehensive set of UAV surveys were performed at 29 sites using visible-colour (VIS-C), multispectral (M.S) and thermal infrared (TIR) cameras covering a variety of soil types, surface and wetness conditions (Paper III). Amongst the three cameras employed, the imagery captured by the TIR camera (at 69% of the sites) was more successful in detecting the drainage signature (i.e. shaded linear features) in comparison to the M.S (59%) and VIS-C (48%) cameras. The key findings of this study were: (1) at some sites the M.S and VIS-C cameras were more effective than TIR camera, (2) prior rainfall events can sometimes have an important impact, hence the timing is crucial, and (3) under most circumstances, farm field operations (i.e. wheel tracks or crop residue) produced similar signature as the drain lines. For the latter, to avoid confusion, knowledge of the drainage installations and farm field operations could be employed to distinguish them.
Moreover, as TIR imagery proved superior when compared to VIS-C, M.S cameras for the drainage pipe detection, the time of the day impact on capturing the TIR imagery was evaluated based on a set of sunrise to sunset surveys (Paper IV). The surveys performed during sunrise/sunset sometimes provided excellent results as it was easier to distinguish the drainage pipes’ signature (i.e. darker shaded linear features) from those caused due to farm field operations (i.e. lighter shaded linear features). However, occasionally problems were encountered in processing (i.e. stitching) the data. Less to no problems occurred for processing the data acquired from late morning through late afternoon surveys. Nevertheless, a similar drainage pipe response was observed as that of farm field operations, therefore impeding their discrimination.
The combined ability of the GPR and UAV imagery was also tested as both the techniques differ in the mode of usage, applicability, and the properties they measure or respond to. Further, the techniques demonstrated their own advantages and drawbacks in relation to drainage mapping. This time, a time-domain GPR system was employed instead of the frequency-domain 3D-GPR discussed earlier. The study was conducted at four sites in the Midwest U.S.A, to compare and contrast both the methods (Paper II). While the UAV surveys covered the entire field area, the GPR data were acquired only on a limited spatial extent preferably in the direction perpendicular to the drain line orientation. In instances, where the UAV imagery partly/completely failed to capture the drainage pipes’ location, GPR proved useful to map them. In instances, where the UAV imagery was successful, GPR acted as a suitable validation technique (i.e. to discriminate the linear features). Furthermore, GPR provided depth information of the drainage pipes and insights on the soil drainage status (i.e. drained/undrained). Nonetheless, there was also a site where the GPR failed to locate the drainage pipes, whereas UAV imagery proved successful in mapping them. Therefore, both the techniques proved complementary in subsurface drainage mapping.
Based on the above studies presented in this dissertation and previous research, it can be ascertained that there is “no silver bullet”, i.e. there is no single technique that can be used across all agricultural fields with different soil types and hydrological conditions, for mapping subsurface drainage systems. Hence, an attempt was made to develop guidelines for carrying out the drainage mapping surveys, especially concerning the sensors (i.e. GPR and UAV imagery) employed in this dissertation and suitable recommendations are provided for their individual and combined usage.
Agricultural subsurface drainage systems provide many agronomic, economic, and environmental benefits. However, if improperly managed they have deleterious effects on the environment as they shorten pathways for the transport of solutes to the aquatic ecosystems. Thus, knowledge of these buried drainage pipes’ location is crucial to comprehend the hydrology and solute dynamics and plan effective mitigation strategies. This knowledge can also benefit drainage water management, a conservation practice that is increasingly being adopted to modulate the water table where a control structure can be retrofitted to the existing drainage system and on the contrary, for the drainage intensification, as it is a typical practice to install the new drainage system in between the existing one. Despite this, the location of the drainage systems is often poorly documented or entirely unknown.
Conventional methods (tile probes and trenching equipment) for finding drainage pipes are invasive, tedious and causes pipe damage. Therefore, there is a growing need to find rapid, effective and non-destructive methods for drainage mapping purpose. Previous research showcases that proximal and remote soil and crop sensing methods may provide a potential alternative solution – amongst which maximum success was achieved by employing a ground penetrating radar (GPR) and aerial imagery captured using different cameras. Recent technological advances led to rapid developments in these techniques. While this constitutes the development of a 3D-GPR capable of scanning an ultra-wideband frequency with an antenna array of wide-area swathe, on another front, the unmanned aerial vehicle (UAV) and their associated cameras have become cheaper than ever making them flexible and feasible to capture aerial imagery. Hence, this dissertation aims at mapping subsurface drainage systems with state-of-the-art proximal and remote sensing methods (i.e. GPR and UAV imagery). A novel magnetic gradiometer (tMag) was also used but only to a limited extent and further research was discontinued owing to the lack of success.
In relation to the stepped-frequency continuous wave 3D-GPR, firstly, at one study site in Denmark, the performance of the system was assessed for different survey configurations (i.e. ground-coupled vs air-coupled) and site conditions (i.e. dry vs wet soil) to discern optimal setup and timing for performing the drainage mapping surveys (Paper V). Out of the combinations tested, a clearer drainage pipe response (i.e. hyperbolic patterns in the vertical profiles) was observed in the survey carried on wet soil with snow cover, however, the ringing noise was substantial at places and prevented the detection of the shallow drainage pipes. The surveys performed on dry soil showed a similar drainage pipe response in both the survey configurations. Later, a thorough assessment of a 3D-GPR was made based on a study conducted at 12 sites in Denmark on a variety of soil types (Paper I). Here, algorithms were developed for calculating global and localized penetration depth (PD). Two different approaches were tested for determining the localized PD. While the first approach was qualitative and helpful to comprehend the attenuation characteristics of the subsurface, the second approach permitted quantification and allowed comparison between different study sites. In both approaches, efforts were made to comprehend the support extended by the electrical conductivity (EC) measured using electromagnetic induction (EMI) instrument to evaluate the performance of the 3D-GPR in finding the drainage pipes. Overall, a high success rate was observed at five amongst the 12 sites visited. At seven sites, the 3D-GPR demonstrated less success owing to a high soil EC (thus limiting the 3D-GPR PD) and the driving direction being parallel to the drainage pipe orientation.
In relation to the UAV imagery, the studies conducted in Midwest U.S.A showcases the swiftly evolving UAV technology and associated cameras as a cost-effective and feasible tool for the drainage mapping purpose. More specifically, to determine the overall feasibility, a comprehensive set of UAV surveys were performed at 29 sites using visible-colour (VIS-C), multispectral (M.S) and thermal infrared (TIR) cameras covering a variety of soil types, surface and wetness conditions (Paper III). Amongst the three cameras employed, the imagery captured by the TIR camera (at 69% of the sites) was more successful in detecting the drainage signature (i.e. shaded linear features) in comparison to the M.S (59%) and VIS-C (48%) cameras. The key findings of this study were: (1) at some sites the M.S and VIS-C cameras were more effective than TIR camera, (2) prior rainfall events can sometimes have an important impact, hence the timing is crucial, and (3) under most circumstances, farm field operations (i.e. wheel tracks or crop residue) produced similar signature as the drain lines. For the latter, to avoid confusion, knowledge of the drainage installations and farm field operations could be employed to distinguish them.
Moreover, as TIR imagery proved superior when compared to VIS-C, M.S cameras for the drainage pipe detection, the time of the day impact on capturing the TIR imagery was evaluated based on a set of sunrise to sunset surveys (Paper IV). The surveys performed during sunrise/sunset sometimes provided excellent results as it was easier to distinguish the drainage pipes’ signature (i.e. darker shaded linear features) from those caused due to farm field operations (i.e. lighter shaded linear features). However, occasionally problems were encountered in processing (i.e. stitching) the data. Less to no problems occurred for processing the data acquired from late morning through late afternoon surveys. Nevertheless, a similar drainage pipe response was observed as that of farm field operations, therefore impeding their discrimination.
The combined ability of the GPR and UAV imagery was also tested as both the techniques differ in the mode of usage, applicability, and the properties they measure or respond to. Further, the techniques demonstrated their own advantages and drawbacks in relation to drainage mapping. This time, a time-domain GPR system was employed instead of the frequency-domain 3D-GPR discussed earlier. The study was conducted at four sites in the Midwest U.S.A, to compare and contrast both the methods (Paper II). While the UAV surveys covered the entire field area, the GPR data were acquired only on a limited spatial extent preferably in the direction perpendicular to the drain line orientation. In instances, where the UAV imagery partly/completely failed to capture the drainage pipes’ location, GPR proved useful to map them. In instances, where the UAV imagery was successful, GPR acted as a suitable validation technique (i.e. to discriminate the linear features). Furthermore, GPR provided depth information of the drainage pipes and insights on the soil drainage status (i.e. drained/undrained). Nonetheless, there was also a site where the GPR failed to locate the drainage pipes, whereas UAV imagery proved successful in mapping them. Therefore, both the techniques proved complementary in subsurface drainage mapping.
Based on the above studies presented in this dissertation and previous research, it can be ascertained that there is “no silver bullet”, i.e. there is no single technique that can be used across all agricultural fields with different soil types and hydrological conditions, for mapping subsurface drainage systems. Hence, an attempt was made to develop guidelines for carrying out the drainage mapping surveys, especially concerning the sensors (i.e. GPR and UAV imagery) employed in this dissertation and suitable recommendations are provided for their individual and combined usage.
| Originalsprog | Engelsk |
|---|
| Udgivelsessted | Aarhus |
|---|---|
| Forlag | Århus Universitet |
| Antal sider | 238 |
| ISBN (Trykt) | 9788793148925 |
| DOI | |
| Status | Udgivet - apr. 2021 |