Inland waters are usually a fundamental resource; they are also essential for ecological preservation and hydrologic regulation. In addition, they are extremely sensitive to a range of stressors operating at global, regional and local scales and thus to environmental changes. For this reason they have been included in the Essential Climate Variables (ECVs) list. In the latest decade, the need for a continuous monitoring of waters quality state was stressed by several project outcome, as well as requested all over the world by directives such as the European Water Frame directive (2010). Earth Observation (EO) techniques for water quality monitoring, have been spreading in the latest decades as a supporting tool to traditional measurements. The physical quantity mainly used for the retrieval of water quality parameters (WQPs), such as chlorophyll-a concentration (chl-a), is the water reflectance. It cannot be directly measured by remote sensors but should be retrieved instead from their imagery through a processing chain, for the removal of useless signal components, including the contribution due to the atmosphere, and other disturbances. Accurate atmospheric correction is a key step to obtain accurate water reflectance signature, and thus accurate WQPs maps. For inland and coastal waters, in addition, this step is made more challenging by: - the presence of optically complex waters, which hinder the use of traditional atmospheric correction methods - the strong impact over the signal of adjacency effect - the variability of aerosol composition, which strongly affects atmospheric and electromagnetic signal interactions. The first objective of this PhD thesis is the evaluation of suitable processing chains for EO optical data, provided by the latest generation satellite sensors (Landsat-8-OLI, L8, and Sentinel-2A/B-MSI, S2A/B, and Sentinel-3A-OLCI, S3A) for water quality monitoring over optically complex waters, focusing on inland waters. In particular, three case study area were selected, characterised by different trophic conditions: oligo-mesotrophic subalpine lakes and eutrophic Mantua lakes in North Italy; meso-eutrophic Lake Trasimeno in central Italy; hypereutrophic Curonian Lagoon, on the Baltic Sea. Several atmospheric correction techniques were applied to be compared and validated through in situ radiometric measurements. Best performing water reflectance products were used to retrieve chl-a concentration maps through bio-optical model, again validated against in situ measurements. As a second objective, phytoplankton trend analysis over selected study area were performed: eutrophication time trend, phytoplankton phenology, intensive algal bloom detection and potential contribution of EO data to WFD monitoring activities were evaluated, exploiting validated, best performing EO products time series. Good performances were obtained by several algorithms especially for more turbid waters, but satisfying performances were obtained also in more oligotrophic lakes. Sensors products inter-comparison produced very good results encouraging merging their data series for both short and long term trend analysis. Merged time series allowed, together with previously validated ENVISAT-MERIS derived maps, to assess seasonal patterns in lakes but also a slight trend towards eutrophication in subalpine lakes, as well as for pelagic area of Lake Trasimeno. For intensive algal bloom detection, latest sensors products showed to be very efficient and useful, in particular being able to catch rapid and spatially delimited phenomena. Finally, in WFD monitoring application it was clear as higher frequency monitoring provided by EO techniques allows to have a more complete characterization of waters state, both in term of space and time, respect to punctual and less frequent in situ measurements, still fundamental for validation and vertical variability characterization.

Le acque interne sono una risorsa fondamentale ed estremamente sensibile ai cambiamenti climatici e per questo sono state inserite nella lista delle Essential Climate Variables. La necessità di un loro monitoraggio è sottolineata da diversi progetti e da direttive come l’europea Water Frame Directive (WFD). Le tecniche di telerilevamento per il monitoraggio delle acque si sono diffuse negli ultimi decenni come uno strumento di supporto alle misure limnologiche. La grandezza fisica usata per derivare le mappe di parametri di qualità delle acque (WQP), come la concentrazione di clorofilla-a (chl-a), è la riflettanza dell’acqua. Questa non può essere misurata direttamente dai sensori ma deve essere ricavata attraverso una catena di processamento, per la rimozione dei componenti del segnale non utili alla stima dei WQP, fra cui il contributo dell’atmosfera. La correzione atmosferica (AC) è un passaggio chiave per la stima accurata della riflettanza dell’acqua e per la produzione di mappe accurate di WQP. Per le acque interne e costiere, questo passaggio è complicato dalla presenza di acque otticamente complesse, che impediscono l’uso dei tradizionali metodi di AC; dal forte impatto sul segnale dell’effetto di adiacenza; dalla variabilità della composizione dell’aerosol atmosferico, che impatta sulle interazioni fra segnale elettromagnetico ed atmosfera. Il primo obiettivo di questa tesi è la valutazione di catene di processamento delle immagini satellitari prodotte dai sensori ottici di ultima generazione (Landsat-8-OLI, Sentinel-2A/B-MSI e Sentinel-3A-OLCI) per il monitoraggio della qualità delle acque, in acque interne otticamente complesse. In particolare, tre casi studio sono stati selezionati, caratterizzati da tre diversi livelli di trofia: i laghi subalpini, oligo-mesotrofici, ed i laghi eutrofici di Mantova; il lago meso-eutrofico Trasimeno; l’iper-eutrofica Laguna dei Curoni, affacciata sul Mar Baltico. Diverse tecniche di correzione atmosferica sono state applicate, confrontate e validate attraverso misure radiometriche effettuate in situ. I prodotti di riflettanza ottenuti, sono stati poi utilizzati per ricavare mappe di chl-a attraverso modelli bio-ottici, anch’essi validati utilizzando misure in situ. Buone performance sono state ottenute da diversi algoritmi, soprattutto per le acque più produttive, ma soddisfacenti sono stati i risultati ottenuti anche nelle acque oligotrofe. Come secondo obiettivo, è stata effettuata un’analisi delle serie storiche, utilizzando le mappe di chl-a, per valutare eventuali trend temporali nei livelli di trofia, nella fenologia stagionale del fitoplancton, ed il contributo apportato dalle tecniche di telerilevamento nelle attività di monitoraggio legate alla WFD, e per l’individuazione di intense fioriture algali. L’inter-comparazione dei prodotti derivati da sensori diversi ha prodotto risultati molto soddisfacenti, incoraggiando l’uso sinergico delle loro serie temporali per le analisi sia di breve che di lungo periodo. Queste serie, unite alla serie storica dei prodotti ENVISAT-MERIS, hanno permesso di individuare i pattern stagionali della concentrazione di chl-a e una tendenza verso livelli di trofia maggiore nei laghi subalpini e nell’area pelagica del lago Trasimeno. Per il monitoraggio di fioriture algali, i prodotti derivati dai sensori di ultima generazione hanno mostrato una buona efficienza nell’individuare fenomeni rapidi e limitatati sia nel tempo che spazialmente. Infine, per le attività dalla WFD, è evidente come, permettendo una maggiore frequenza, i prodotti satellitari diano la possibilità di meglio caratterizzare in maniera completa lo stato delle acque, sia in termini di tempo che di spazio, rispetto alle misure tradizionali, puntuali e meno frequenti, le quali rimangono fondamentali, non solo per la validazione dei prodotti satellitari, ma anche per la descrizione della variabilità lungo la colonna d’acqua.

(2019). Processing and analysis of latest generation satellite data for monitoring optically complex waters. (Tesi di dottorato, Università degli Studi di Milano-Bicocca, 2019).

Processing and analysis of latest generation satellite data for monitoring optically complex waters

CAZZANIGA, ILARIA
2019

Abstract

Inland waters are usually a fundamental resource; they are also essential for ecological preservation and hydrologic regulation. In addition, they are extremely sensitive to a range of stressors operating at global, regional and local scales and thus to environmental changes. For this reason they have been included in the Essential Climate Variables (ECVs) list. In the latest decade, the need for a continuous monitoring of waters quality state was stressed by several project outcome, as well as requested all over the world by directives such as the European Water Frame directive (2010). Earth Observation (EO) techniques for water quality monitoring, have been spreading in the latest decades as a supporting tool to traditional measurements. The physical quantity mainly used for the retrieval of water quality parameters (WQPs), such as chlorophyll-a concentration (chl-a), is the water reflectance. It cannot be directly measured by remote sensors but should be retrieved instead from their imagery through a processing chain, for the removal of useless signal components, including the contribution due to the atmosphere, and other disturbances. Accurate atmospheric correction is a key step to obtain accurate water reflectance signature, and thus accurate WQPs maps. For inland and coastal waters, in addition, this step is made more challenging by: - the presence of optically complex waters, which hinder the use of traditional atmospheric correction methods - the strong impact over the signal of adjacency effect - the variability of aerosol composition, which strongly affects atmospheric and electromagnetic signal interactions. The first objective of this PhD thesis is the evaluation of suitable processing chains for EO optical data, provided by the latest generation satellite sensors (Landsat-8-OLI, L8, and Sentinel-2A/B-MSI, S2A/B, and Sentinel-3A-OLCI, S3A) for water quality monitoring over optically complex waters, focusing on inland waters. In particular, three case study area were selected, characterised by different trophic conditions: oligo-mesotrophic subalpine lakes and eutrophic Mantua lakes in North Italy; meso-eutrophic Lake Trasimeno in central Italy; hypereutrophic Curonian Lagoon, on the Baltic Sea. Several atmospheric correction techniques were applied to be compared and validated through in situ radiometric measurements. Best performing water reflectance products were used to retrieve chl-a concentration maps through bio-optical model, again validated against in situ measurements. As a second objective, phytoplankton trend analysis over selected study area were performed: eutrophication time trend, phytoplankton phenology, intensive algal bloom detection and potential contribution of EO data to WFD monitoring activities were evaluated, exploiting validated, best performing EO products time series. Good performances were obtained by several algorithms especially for more turbid waters, but satisfying performances were obtained also in more oligotrophic lakes. Sensors products inter-comparison produced very good results encouraging merging their data series for both short and long term trend analysis. Merged time series allowed, together with previously validated ENVISAT-MERIS derived maps, to assess seasonal patterns in lakes but also a slight trend towards eutrophication in subalpine lakes, as well as for pelagic area of Lake Trasimeno. For intensive algal bloom detection, latest sensors products showed to be very efficient and useful, in particular being able to catch rapid and spatially delimited phenomena. Finally, in WFD monitoring application it was clear as higher frequency monitoring provided by EO techniques allows to have a more complete characterization of waters state, both in term of space and time, respect to punctual and less frequent in situ measurements, still fundamental for validation and vertical variability characterization.
COLOMBO, ROBERTO
GIARDINO, CLAUDIA
Telerilevamento; fitoplancton; acque interne; laghi; clorofilla
remote sensing; phytoplankton; inland waters; lakes; clorofilla
GEO/10 - GEOFISICA DELLA TERRA SOLIDA
English
20-feb-2019
SCIENZE CHIMICHE, GEOLOGICHE E AMBIENTALI - 94R
31
2017/2018
open
(2019). Processing and analysis of latest generation satellite data for monitoring optically complex waters. (Tesi di dottorato, Università degli Studi di Milano-Bicocca, 2019).
File in questo prodotto:
File Dimensione Formato  
phd_unimib_809088.pdf

accesso aperto

Descrizione: tesi di dottorato
Dimensione 8.84 MB
Formato Adobe PDF
8.84 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10281/241143
Citazioni
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
Social impact