Kelp forests are highly productive three-dimensional marine ecosystems that provide valuable ecosystem services globally. Along the coast of British Columbia, Macrocystis pyrifera and Nereocystis luetkeana are two key species that form surface-canopies that are vulnerable to both biotic and abiotic drivers; making it imperative to monitor and understand whether these ecosystems are changing in the face of climate change. The monitoring of kelp forests is commonly enhanced by use of remote sensing, which allows researchers to survey large portions of the coast where it would otherwise be difficult to collect data, and to use archived imagery for comparisons of historic and contemporary kelp forest trends. Generally, the remote sensing of kelp surface-canopy relies on differences in the high near-infrared (NIR; 700-1000 nm) signal of kelp and the low NIR signal of water. However, kelp surface-canopy reflectance signals can be affected by submergence under water, caused by oceanographic features like tides and currents, or simply due to differences in the morphology and buoyancy of kelp canopy structures. This submersion may cause uncertainties when estimating the surface-canopy area of kelp beds in remote sensing imagery. This research aims to understand the effects of submersion on the remote sensing of kelp surface-canopy. To address our goal, (i) Nereocystis canopy structures (bulb and blade) were submerged while collecting above-water hyperspectral measurements. The hyperspectral data into the bandwidths of high-resolution multispectral aerial and space-borne sensors and vegetation indices were calculated to understand the kelp detection limits when using shorter red-edge wavelengths (RE; 690-750 nm) instead of the longer NIR wavelengths. The results showed that submerged kelp can be detected deeper in the water column using shorter RE wavelengths compared to the more commonly used NIR wavelengths. Further, (ii) in situ hyperspectral data were also collected for the different surface-canopy structures and compared with UAV imagery, which showed that the buoyancy of the kelp canopy structures at the surface affected the relative magnitude of reflectance in both the RE and NIR and supported the findings of the submersion experiment. The total surface-canopy area derived from classification with both RE and NIR vegetation indices were compared in the UAV imagery, and the RE index detected roughly 18% more kelp than the NIR index, with no differences seen between Macrocystis and Nereocystis, or between high and low tide in beds larger than 150m2. Finally, (iii) to understand how submersion by tides and currents affect the ability to estimate surface-canopy area for both Macrocystis and Nereocystis, surface-canopy area was derived from multispectral unoccupied aerial vehicle (UAV) imagery and compared with in situ tide and current data, which showed that surface-canopy area had a strong negative linear relationship with tidal height at all sites regardless of species. Macrocystis occupied sites where currents were low (<10cm/s) and did not affect the surface-canopy. Therefore, the extent of all Macrocystis beds decreased at a similar rate over their tidal range (22.7 + 2.8%/m). Nereocystis beds occupied a wider range of current speeds (0.0 - 19.0 cm/s), and at sites with high current speeds (> 10 cm/s) increasing current and tidal height decreased surface-canopy area simultaneously, resulting in both a higher and more variable rate of decrease (30.5 + 9.1%/m) with increasing tidal height than Macrocystis. Together, this thesis addressed critical questions related to the effects of kelp submersion on the remote sensing of surface-canopy forming kelps, and we provide recommendation for remote sensors who wish to minimize errors when using remote sensing to map kelp forests. / Graduate

Conservation of elasmobranch populations is often inhibited by a lack of data, particularly in understudied regions like the Red Sea. Survey efforts in this region have been infrequent and often highly localized. Establishing a broad baseline for elasmobranch diversity and abundance along the Saudi Arabian Red Sea coast could inform both conservation efforts and a nascent ecotourism industry. In this thesis, I describe a pilot study comparing biodiversity data from baited remote underwater video stations (BRUVS), unoccupied aerial vehicle surveys (UAVs), and eDNA sequencing at five islands in the Farasan Banks region of the Saudi Arabian Red Sea. Estimates of relative abundance were also compared between the BRUVS and UAVs. Each method identified species missed by the other two, but all three techniques exhibited clear habitat- and taxa-specific biases. I was able to identify key concerns for each approach that need to be addressed before large-scale implementation. If carefully planned and executed well, a full assessment of the Saudi Arabian coastline could establish a true baseline for shallow water elasmobranchs in the eastern Red Sea. Informing best conservation practices and identifying potential ecological attractions in accordance the environmental and economic goals of Saudi Arabia’s Vision 2030.

La télédétection des forêts est devenue de plus en plus accessible grâce à l'utilisation de véhicules aériens inoccupés (UAV) et à l'apprentissage profond, ce qui permet d'obtenir des images répétées à haute résolution et d’observer les changements phénologiques à des échelles spatiales et temporelles plus importantes. Dans les forêts tempérées, à l'automne, la sénescence des feuilles se produit lorsque les feuilles changent de couleur et tombent. Cependant, l'influence de la sénescence foliaire sur la segmentation des espèces d'arbres à l'aide d'un réseau neuronal convolutif (CNN) n'a pas encore été évaluée. Nous avons acquis de l’imagerie haute résolution par UAV au-dessus d’une forêt tempérée au Québec à sept reprises entre mai et octobre 2021. Nous avons segmenté et identifié 23 000 couronnes d'arbres de 14 classes différentes pour entraîner et valider un CNN pour chaque acquisition d'imagerie. La meilleure segmentation (F1-score le plus élevé) était au début de la coloration des feuilles (début septembre) et le F1-score le plus bas au pic de la coloration automnale (début octobre). La chronologie de la sénescence varie considérablement d’une espèce à l’autre et au sein d’une même espèce, ce qui entraîne une grande variabilité du signal télédétecté. Les espèces d'arbres à feuilles caduques et à feuilles persistantes qui présentaient des traits distinctifs et moins variables dans le temps entre les individus ont été mieux classées. Bien que la segmentation des arbres dans une forêt hétérogène demeure un défi, l'imagerie UAV et l'apprentissage profond démontrent un grand potentiel pour la cartographie des espèces d'arbres. Les résultats obtenus dans une forêt tempérée où la couleur des feuilles change fortement pendant la sénescence automnale montrent que la meilleure performance pour la segmentation des espèces d'arbres se produit au début de ce changement de couleur. / Remote sensing of forests has become increasingly accessible with the use of unoccupied aerial vehicles (UAV), along with deep learning, allowing for repeated high-resolution imagery and the capturing of phenological changes at larger spatial and temporal scales. In temperate forests during autumn, leaf senescence occurs when leaves change colour and drop. However, the influence of leaf senescence in temperate forests on tree species segmentation using a Convolutional Neural Network (CNN) has not yet been evaluated. Here, we acquired high-resolution UAV imagery over a temperate forest in Quebec, Canada on seven occasions between May and October 2021. We segmented and labelled 23,000 tree crowns from 14 different classes to train and validate a CNN for each imagery acquisition. The CNN-based segmentation showed the highest F1-score (0.72) at the start of leaf colouring in early September and the lowest F1-score (0.61) at peak fall colouring in early October. The timing of the events occurring during senescence, such as leaf colouring and leaf fall, varied substantially between and within species and according to environmental conditions, leading to higher variability in the remotely sensed signal. Deciduous and evergreen tree species that presented distinctive and less temporally-variable traits between individuals were better classified. While tree segmentation in a heterogenous forest remains challenging, UAV imagery and deep learning show high potential in mapping tree species. Our results from a temperate forest with strong leaf colour changes during autumn senescence show that the best performance for tree species segmentation occurs at the onset of this colour change.