Les particules en suspension dans l'eau de mer incluent les sédiments, le phytoplancton, le zooplancton, les bactéries, les virus et des détritus. Ces particules sont communément appelés matière en suspension (MES). Dans les eaux côtières, la MES peut parcourir de longues distances et être transportée verticalement à travers la colonne d'eau sous l'effet des vents et des marées favorisant les processus d'advection et de resuspension. Ceci implique une large variabilité spatio-temporelle de MES et quasiment impossible à reconstituer à travers les mesures traditionnelles des concentrations de MES [MES], par filtration de l'eau de mer à bord de bateaux. La [MES] peut être obtenue à partir de capteurs optiques enregistrant la diffusion et déployés soit de manière in-situ, soit à partir d'un satellite dans l'espace. Depuis la fin des années 70, par exemple, les satellites "couleur de l'eau" permettent d'établir des cartes de [MES] globales. La fréquence d'une image par jour pour la mer di Nord de ces capteurs polaires représente un obstacle non négligeable pour l'étude de variabilité de la [MES] dans les eaux côtières où la marée et les vents engendrent des variations rapides au cours de la journée. Cette limitation est d'autant plus importante pour les régions avec une couverture nuageuse fréquente. Les méthodes in-situ à partir d'un navire autonome ou d'une plateforme amarrée permettent d'enregistrer des données en continu mais leur couverture spatiale reste néanmoins limitée. Ce travail a pour objectif de mettre en avant les techniques de mesures in-situ et satellite de la [MES] en se concentrant principalement sur deux points. Premièrement, d'acquérir une meilleure connaissance de la variabilité de la relation entre la [MES] et la lumière diffuse, et deuxièmement, d'établir des cartes de [MES] dans la mer du Nord avec le capteur géostationnaire météorologique Européen (SEVIRI) qui donne des images chaque 15 minutes.La variabilité de la relation entre la [MES] et la lumière diffuse est étudiée à l'aide d'une banque de données in-situ. Nous démontrons que la [MES] est le mieux estimée à partir des mesures dans l'intervalle rouge du spectre de lumière rétro-diffuse. Par ailleurs, la relation entre la [MES] et la rétrodiffusion est gouvernée par la composition organique/inorganique des particules, ce qui représente des possibilités d'amélioration pour les algorithmes d'estimation de [MES] à partir de la couleur de l'eau. Nous démontrons aussi qu'avec SEVIRI il est possible d'estimer la [MES], la turbidité et le coefficient d'atténuation, deux variables étroitement liées à la [MES], avec généralement une bonne précision. Bien qu'il y ait d'importantes incertitudes dans les eaux claires, cette réussite est remarquable pour un capteur météorologique initialement conçu pour le suivi des nuages et des masses glaciaires, cibles beaucoup plus brillantes que la mer! Ce travail démontre pour la première fois que la variabilité de la [MES] à l'échelle temporelle des marées dans les eaux côtières au sud de la mer du Nord peut être capturée et mesurée par le biais de la télédétection de la couleur de l'eau ; ce qui ouvre des opportunités pour le monitoring de la turbidité et pour la modélisation des écosystèmes. Le premier capteur géostationnaire couleur de l'eau a été lancé en juin 2012, donnant des images multispectrale des eaux coréennes chaque heure. D'autres capteurs vont probablement suivre dans l'avenir, couvrant le reste des eaux du globe. Ce travail nous permet donc de préparer, de façon optimale, l'arrivée de ces capteurs qui vont révolutionner l'océanographie optique. / Particles suspended in seawater include sediments, phytoplankton, zooplankton, bacteria, viruses, and detritus, and are collectively referred to as suspended particulate matter, SPM. In coastal waters, SPM is transported over long distances and in the water column by biological, tide or wind-driven advection and resuspension processes, thus varying strongly in time and space. These strong dynamics challenge the traditional measurement of the concentration of SPM, [SPM], through filtration of seawater sampled from ships. Estimation of [SPM] from sensors recording optical scattering allows to cover larger temporal or spatial scales. So called ocean colour satelittes, for example, have been used for the mapping of [SPM] on a global scale since the late 1970s. These polar-orbiting satellites typically provide one image per day forthe North Sea area. However, the sampling frequency of these satellites is a serious limitation in coastal waters where [SPM] changes rapidly during the day due to tides and winds.Optical instruments installed on moored platforms or on under-water vehicles can be operated continuously, but their spatial coverage is limited. This work aims to advance in situ and space-based optical techniques for [SPM] retrieval by investigating the natural variability in the relationship between [SPM] and light scattering by particles and by investigating whether the European geostationary meteorological SEVIRI sensor, which provides imagery every 15 minutes, can be used for the mapping of [SPM] in the southern North Sea. Based on an extensive in situ dataset, we show that [SPM] is best estimated from red light scattered in the back directions (backscattering). Moreover, the relationship between [SPM]] and particulate backscattering is driven by the organic/inorganic composition of suspended particles, offering opportunities to improve [SPM] retrieval algorithms. We also show that SEVIRI successfully retrieves [SPM] and related parameters such as turbidity and the vertical light attenuation coefficient in turbid waters. Even though uncertainties are considerable in clear waters, this is a remarkable result for a meteorological sensor designed to monitor clouds and ice, much brighter targets than the sea! On cloud free days, tidal variability of [SPM] can now be resolved by remote sensing for the first time, offering new opportunities for monitoring of turbidity and ecosystem modelling. In June 2010 the first geostationary ocean colour sensor was launched into space which provides hourly multispectral imagery of Korean waters. Other geostationary ocean colour sensors are likely to become operational in the (near?) future over the rest of the world's sea. This work allows us to maximally prepare for the coming of geostationary ocean colour satellites, which are expected to revolutionize optical oceanography. / De in zeewater aanwezige zwevende materie zoals sedimenten, fytoplankton, zooplankton, bacteriën, virussen en detritus, worden collectief "suspended particulate matter" (SPM) genoemd. In kustwateren worden deze deeltjes over lange afstanden en in de waterkolom getransporteerd door biologische processen of wind- of getijdenwerking, waardoor SPM sterk varieert in ruimte en tijd. Door deze sterke dynamiek wordt de traditionele bemonstering van de concentratie van SPM, [SPM], door middel van filtratie van zeewaterstalen aan boord van schepen ontoereikend. Optische technieken die gebruik maken van de lichtverstriioongseigenschappen van SPM bieden een gebieds- of tijdsdekkend alternatief. Zogenaamde "ocean colour" satellieten bijvoorbeeld leveren beelden van o.a. [SPM] aan het zeeoppervlak op globale schaal sinds eind 1970, met een frequantie van één beeld per dag voor de Noordzee. Deze frequentie is echter onvoldoende in onze kustwateren waar [SPM] drastisch kan veranderen in enkele uren tijd. Optische instrumenten aan boord vann schepen of op onderwatervoertuigen kunnen continu meten, maar de gebiedsdekking is deperkt. Dit werk heeft tot doel de lichtverstriioongseigenschappen van SPM te karakterizeren en te onderzoeken of de Europese geostationaire weersatelliet, die elk kwartier een beeld geeft, kan worden gebruikt voor de kartering van [SPM] in de zuidelijke Noordzee. Op basis van een grote dataset van in situ metingen tonen wij aan dat [SPM] het nauwkeurigst kan worden bepaald door de meting van de verstrooiing van rood licht in achterwaartse richtingen (terugverstrooiing). Bovendien blijkt de relatie tussen [SPM] en terugverstrooiing afhankelijk van de organische-anorganische samenstelling van zwenvende stof, wat mogelijkhenden biedt tot het verfijnen van teledetectiealgoritmen voor [SPM]. Voorts tonen woj aan dat de Europese weersatelliet, SEVIRI, successvol kan worden aangewend voor de kartering van [SPM] en gerelateerde parameters zoals troebelheid en lichtdemping in de waterkolom. Hoewel met grote meetonzekerheid in klaar water toch een opmerkelijk resultaat voor een sensor die ontworpen werd voor detectie van wolken en ijs! Op wolkenvrije dagen wordt hierdoor de getijdendynamiek van [SPM] in de zuidelijke Noordzee voor het eerst detecteerbaar vanuit de ruimte, wat nieuwe mogelijkheden biedt voor de monitoring van waterkwaliteit en verbetering van ecosysteellodellen. Sinds juni 2010 is de eerste geostationaire ocean colour satelliet een feit : elk uur een multispectraal beeld van Koreaanse wateren. Vermoedelijk zullen er in de (nabije?) toekomst meer volgen over Europa en Amerika. Dit werk laat toe ons maximaal voor te bereiden op te komst van zo'n satellieten, waarvan verwacht wordt dat zij een nieuwe revolutie in optische oceanografie zullen ontketenen.

Hyperspectral imaging has become one of the most popular technologies in plant phenotyping because it can efficiently and accurately predict numerous plant physiological features such as plant biomass, leaf moisture content, and chlorophyll content. Various hyperspectral imaging systems have been deployed in both greenhouse and field phenotyping activities. However, the hyperspectral imaging quality is severely affected by the continuously changing environmental conditions such as cloud cover, temperature and wind speed that induce noise in plant spectral data. Eliminating these environmental effects to improve imaging quality is critically important. In this thesis, two approaches were taken to address the imaging noise issue in greenhouse and field separately. First, a computational simulation model was built to simulate the greenhouse microclimate changes (such as the temperature and radiation distributions) through a 24-hour cycle in a research greenhouse. The simulated results were used to optimize the movement of an automated conveyor in the greenhouse: the plants were shuffled with the conveyor system with optimized frequency and distance to provide uniform growing conditions such as temperature and lighting intensity for each individual plant. The results showed the variance of the plants’ phenotyping feature measurements decreased significantly (i.e., by up to 83% in plant canopy size) in this conveyor greenhouse. Secondly, the environmental effects (i.e., sun radiation) on <a>aerial </a>hyperspectral images in field plant phenotyping were investigated and modeled. <a>An artificial neural network (ANN) method was proposed to model the relationship between the image variation and environmental changes. Before the 2019 field test, a gantry system was designed and constructed to repeatedly collect time-series hyperspectral images with 2.5 minutes intervals of the corn plants under varying environmental conditions, which included sun radiation, solar zenith angle, diurnal time, humidity, temperature and wind speed. Over 8,000 hyperspectral images of </a>corn (<i>Zea mays </i>L.) were collected with synchronized environmental data throughout the 2019 growing season. The models trained with the proposed ANN method were able to accurately predict the variations in imaging results (i.e., 82.3% for NDVI) caused by the changing environments. Thus, the ANN method can be used by remote sensing professionals to adjust or correct raw imaging data for changing environments to improve plant characterization.

Abstract The diurnal cycle is a dominant mode of sea surface temperature (SST) variability in trop-ical oceans, that influences air-sea interaction and climate processes. Diurnal variability of SST generally ranges from ~0.1 to 2.0◦C and is controlled by atmospheric fluxes of heat and momentum. In the present study, the response of intraseasonal variability (ISV) of SST in the Bay of Bengal (BoB) to diurnal atmospheric forcings, during the summer monsoon of 2007, has been examined using an Ocean General Circulation Model (OGCM). The model is based on the Modular Ocean Model Version 4 (MOM4p0), having a horizontal resolution of 0.25◦ and 40 vertical levels, with a fine resolution of 5 m in the upper 60 m. Numerical experiments were conducted by forcing the model with daily and hourly atmospheric forcings to examine the SST-ISV modulation with the diurnal cycle. Additional experiments were performed to determine the relative role of diurnal cycle in solar radiation and winds on SST and mixed layer depth (MLD). Since salinity, which is decisive in SST variability, varies meridionally in the BoB, two locations were selected for analyses: one in the northern bay at 89◦E, 19◦N where salinity is lower and the other in the southern bay at 90◦E, 8◦N where salinity is higher, as well as observations are available from Research Moored Array for African-Asian-Australian Monsoon Analysis and Prediction (RAMA) buoy for comparision with model simulation. Diurnal atmospheric forcings modify SST-ISV in both southern and northern bay. SST-ISV in the southern bay, is dominantly controlled by the diurnal cycle of insolation, while in the northern bay, diurnal cycle of insolation and winds have comparable contribution. Diurnal cycle enhanced the amplitude of 3 selected intraseasonal events in the southern bay and 3 out of the 6 events in the northern bay, during the study period. In the southern bay, simulated SST variability with hourly forcing was closer to the observations from RAMA, implying that incorporating the diurnal cycle in model forcing rectifies SST-ISV. Moreover, SST obtained with diurnal forcing consists of additional fluctuations at higher frequencies within and in between intraseasonal events; such fluctuations are absent with daily forcing. The diurnal variability of SST is significant during the warming phase of intraseasonal events and reduces during the cooling phase. Diurnal amplitude of SST decreases with depth; depth dependence also being larger during the warming phase. SST-ISV modulation with diurnal forcing results from the diurnal cycle of upper ocean heat fluxes and vertical mixing. Diurnal warming and cooling result in a net gain or loss of heat in the mixed layer after a day’s cycle. When the retention (loss) of heat in the mixed layer increases with diurnal forcing during the warming (cooling) phase of intraseasonal events, the daily mean SST rise (fall) becomes higher, amplifying the intraseasonal warming (cooling). In the southern bay, SST-ISV amplification is mainly controlled by the diurnal variability of MLD, which modifies the heat fluxes. Increased intraseasonal warming with diurnal forcing results from the increase in radiative heating, due to the shoaling of the daytime mixed layer. Amplified intraseasonal cooling is dominantly con-trolled by the strengthening of sub-surface processes, due to the nocturnal deepening of mixed layer and increased temperature gradients below the mixed layer. In the northern bay, SST-ISV modulation with diurnal forcing is not as large as that in the southern bay. The mean increase in SST-ISV amplitudes with diurnal forcing is ~0.16◦C in the southern bay, while it is only ~0.03◦C in the northern bay. Reduced response of SST-ISV to diurnal forcings in the northern bay is related to the weaker diurnal variability of MLD. Salinity stratification limits diurnal variability of mixed layer in the northern bay, unlike in the southern bay. The seasonal (June - September) mean diurnal amplitude of MLD is ~15 m in the southern bay, while it is reduced to ~1.5 m in the northern bay. Diurnal variability of MLD, spanning only a few meters is not sufficient to create large modifications in mixed layer heat fluxes and SST-ISV in the northern bay. The vertical resolution of the model limits the shallowing of mixed layer to 7.5 m, thus restricting the diurnal variability of simulated MLD.