Drone-Based Ecohydraulic Signatures of Fully-Vegetated Ditches: Real-Scale Experimental Analysis ^†

Abstract

The prediction of the ecohydraulic and ecohydrological structures of vegetated flows is extremely important; particularly because of the well-known Climate Change and Global Warming implications. This study aims at monitoring the real-scale Ecohydrodynamics of vegetated ditches hardly covered by riparian stands, through the analysis of their reflectance signature. A monitoring system composed of a cruising drone and a thermographic camera was employed in the present experimental study. Once the main average and turbulence traits were measured directly in the field, the correlation between the riparian plant’s Leaf Area Index (LAI) and their thermal properties was studied to assess the impact of drone-based thermometry methods and techniques on the predictions of the main real-scale flow trends of vegetated waterways.

1. Introduction

The real-scale analysis of the ecohydraulic patterns exhibited by vegetated water flows is highly affected by the presence of riparian patches [1,2,3]. Specifically, the ecohydrodynamic behavior of vegetation in rivers depends on its morphometrical and ecological traits [4,5,6,7,8]. In the last decade, the potential of drones in monitoring both natural and urban areas has improved considerably, thanks to their remote sensing capabilities, especially when dealing with vegetated water systems almost worldwide [9,10,11,12].

The present experimental research aims at evaluating the potential of image processing techniques of drone-based thermographic acquisition in assessing the actual eco-hydrodynamic traits of a vegetated ditch colonized by mature giant reed (Arundo donax L.) plants [13,14,15,16,17]. In the present study, the field-scale Leaf Area Index (LAI), representing the canopy biomass distribution, and drone-based temperature T (°C) were evaluated at ten 5 m wide experimental cross-sections identified along the examined vegetated ditch, located 3 m apart.

2. Materials and Methods

Study Site

As indicated in Figure 1, the experimental water body studied in the present research is a vegetated open channel located in Nola (Campania region, southern Italy) hardly covered by rigid riparian vegetation, consisting of giant reed plants.

The drone and the corresponding controller selected for this experimental ecohydraulic study are illustrated in Figure 2a and Figure 2b, respectively.

As shown in Figure 2a and Figure 2b, the drone and the thermographic camera selected for the present study are the DJI^® M 600 pro hexacopter and TeleDYNE^® FLiR VUE pro R devices, respectively.

In

Table 1. Most relevant technical characteristics of the thermographic camera sensors employed in the present research.

Technical Characteristics	Values
Weight	110 g
Focal length	13 mm
Resolution Accuracy Band range Thermal sensitivity	336 × 256 ±0.5 °C 8.0 μm–12.0 μm 30 mK
Frequency	20 Hz

, the main technical features of the thermographic camera sensors selected for this work are detailed.

3. Results and Discussion

The map of experimental temperature values associated with the examined vegetated ditch is reported in Figure 3.

In Figure 4 is shown the linear correlation obtained between the experimental values of temperature T (°C) and ground-based Leaf Area Index (LAI) corresponding to ten randomly-distributed cross-sections monitored along the vegetated ditch.

The linear regression represented in Figure 4 is quantitatively represented by the following equation:

LAI = 0.03 T + 1.78.

(1)

The previous equation is characterized by a coefficient of determination R² pair to 0.72, indicating a good level of correlation between T (°C) and LAI at ten cross-sections examined in this work. This outcome represents the first advance in the thermographic-based experimental analysis of the ecohydraulic flow dynamic behaviour of vegetated water flows.

The predictive performance of the experimental methodology adopted in the present study can be examined in more detail in terms of uncertainty analysis, by taking into account the most robust statistical models, widely employed in geoscience and hydraulic modeling research [18,19,20,21,22,23,24,25,26].

4. Conclusions

The main findings of this research represent a first advance in the cruising of experimental ecohydraulic research focused on the protection of natural resources along vegetated water streams at real scale [27,28,29,30,31,32,33,34].

Deep learning and artificial neural network applications [35,36,37,38,39,40,41] can be implemented in the processing of the ecohydraulic datasets retrieved in the present study to simulate the bio-mechanical behavior of the examined riparian plants over time in multiple hydrodynamic conditions [42,43,44,45,46,47,48].