Fluvial processes monitoring and management are effective in the assessment and mitigation of flood risk, in the design of sustainable restoration projects and in the protection of fluvial ecosystem biodiversity and ecosystems services. They play, therefore, a significant role in challenging themes such as climate change, emergency management, and security. Fluvial geomorphological surveys are useful to support a sustainable river management and a comprehensive understanding of the morphodynamics of river systems. A key variable, among multiple factors, to be studied is bed grain size distribution. Grain size mapping and its monitoring overtime is necessary to understand geomorphological processes. Sediment dimensions and their pattern of change along rivers are important for mechanical and fluid dynamic properties of sediments, regulating channel morphology and hydraulics, as well as fluvial ecosystems functioning (Surian, 2001). Nowadays, we are aware of the importance of sampling and studying river catchments as holistic systems in order to support a successful river management. Without a large spatial and temporal understanding, river management cannot be fully effective and sustainable. In response to these requirements remotely sensed data are fundamental for hydromorphological assessment and monitoring, providing continuous data over entire catchments and allowing for a comprehensive river management. In the last two decades, Remote Sensing (RS) technology in river sciences has dramatically increased opening up unprecedented possibilities for river science and management. However, due to the high cost of RS data acquisition and the low spatial resolution available, field-based approaches are still dominating in river surveys and their prerequisites limit their application at the network scale. Recently, the most ambitious Earth observation programme to date, Copernicus, headed by the European Commission (EC) in partnership with the European Space Agency (ESA), has opened up unparalleled possibilities for environmental purposes and monitoring. ESA has developed a new family of missions called Sentinels that supply accurate, timely and easily accessible information. This innovation in RS satellite technology will provide an unparalleled amount of free data available to be investigated. The purpose of this work is to explore the potential of Sentinel 2 data on grain size mapping. Finding the relationship between multispectral information and grain size percentiles and assess the maximum attainable accuracy, creates the possibility to gain important information about broad classes of sediment along big rivers and their changes over time. Moreover, beside testing satellite potential, the broad scope is to find an efficient and repeatable methodology for grain size mapping. The solution we proposed, is the fusion of satellite and Unmanned Aerial Systems (sUAS) remote sensed imagery. The former provides the spectral information, the latter is useful to extract, by photosieving technique, ground truth grain size data, needed for validation. We tested this approach on the major Italian river, Po, in the north of Italy and we identified a functional model for describing the relationship between grain size percentiles and multispectral data. Results show that there is a significant relationship between satellite data and grain size percentiles and this represents a major advance in our current ability to characterize fluvial habitats and processes along major river systems, thanks to the characteristics of Sentinel 2 data, nowadays free available worldwide with a time frequency of about 5 days. The proposed methodology will allow for an objective and rapid grain size mapping, easily transferable to any large river areas, opening up broad perspectives for future fluvial survey practices.
La gestione dei sistemi fluviali è un elemento cruciale nel quadro di sfide attuali quali i cambiamenti climatici, la gestione delle emergenze e della sicurezza del territorio. Per meglio comprendere tali sistemi, la geomorfologia fluviale gioca un ruolo rilevante e fornisce strumenti fondamentali. Una variabile importante da considerare è la granulometria dei sedimenti che costituiscono il letto del fiume e le barre laterali, depositate lungo il suo corso. La dimensione dei sedimenti e la sua variazione lungo il corso del fiume, infatti, è espressione e al contempo un fattore, di processi fluviali, con importanti implicazioni sia sulla morfologia, l’ idraulica (resistenza al flusso, trasporto di sedimenti), che sul funzionamento degli ecosistemi fluviali. Per avere una gestione delle risorse naturali il più possibile esaustiva e sostenibile, è fondamentale considerare e studiare i sistemi ambientali, ed in particolare quello fluviale, in maniera olistica e su larga scala. Le tecnologie di telerilevamento rispondono a queste esigenze e vengono per questo ampiamente utilizzate, fornendo dati a larga scala e con alta frequenza, grazie alla loro alta copertura spaziale e risoluzione temporale. Negli ultimi decenni, il telerilevamento (Remote Sensing) ha trovato una grande applicazione anche nell’ ambito della ricerca fluviale, aprendo nuove possibilità. Tuttavia, a causa sia dei costi di tale tecnologia che spesso della limitata risoluzione spaziale delle immagini disponibili, i campionamenti in situ, sono ancora necessary e diffusi. Questo comporta diversi limiti sia in termini di tempo e di risorse necessari che in termini di area investigabile. Recentemente, il programma europeo Copernicus, ha gettato le basi per un nuovo periodo nell’ uso dei dati satellitari. L’ Unione Europea (EU) in partnership con L’Agenzia Spaziale Europea (ESA), ha sviluppato una nuova famiglia di satelliti, i Sentinels, che forniscono dati accurati, puntuali e facilmente scaricabili, aprendo nuove prospettive nel campo del monitoraggio della terra da remoto. Un grande punto di forza dei dati Sentinel è che sono disponibili gratuitamente e questo consente la fruibilità di un quantitativo di dati senza precedenti. L’obiettivo di questa tesi è quello di indagare le potenzialità dei dati satellitari registrati da Sentinel 2 nel discriminare classi di sedimenti fluviali. Vogliamo studiare la relazione, e la sua accuratezza, tra i dati multispettrali satellitari e i diametri percentili di diverse classi granulometriche (ciottoli, ghiaia, sabbia). Tale relazione permetterebbe di mappare su grandi fiumi, a scala globale, la distribuzione granulometrica dei sedimenti, rivoluzionando le attuali metodologie disponibili. Il metodo di lavoro proposto prevede l’utilizzo integrato di due tecnologie: satellite e drone. Il primo fornisce le informazioni spettrali, il secondo è utile per ricavare le verità a terra delle dimensioni dei grani. Per verificare la relazione tra queste due variabili si utilizza un modello di regressione multivariata. Nella metodologia si procede con l’acquisizione delle immagini da drone nelle zone di interesse e si valutano, per ogni immagine acquisita, i valori dei diametri percentili che sono le variabili di risposta con cui verrà costruito il modello. Parallelamente si scaricano le immagini satellitari, da cui si ricavano i valori di riflettanza, che costituiscono i predittori del modello di regressione lineare multivariato. La metodologia descritta è stata applicata al più grande fiume italiano, il fiume Po. I risultati mostrano una significativa correlazione tra i dati satellitari e i diametri percentili. Tale relazione apre promettenti potenzialità per avanzare nelle attuali metodologie di mappatura delle granulometrie dei sedimenti e, grazie alle caratteristiche tecniche di Sentinel 2, i cui dati sono disponibili gratuitamente, su scala globale, con una frequenza di circa 5 giorni, può essere facilmente replicata su altri grandi fiumi, a scala globale.
Overall grain size mapping from Sentinel 2 images
MARCHETTI, GIULIA
2016/2017
Abstract
Fluvial processes monitoring and management are effective in the assessment and mitigation of flood risk, in the design of sustainable restoration projects and in the protection of fluvial ecosystem biodiversity and ecosystems services. They play, therefore, a significant role in challenging themes such as climate change, emergency management, and security. Fluvial geomorphological surveys are useful to support a sustainable river management and a comprehensive understanding of the morphodynamics of river systems. A key variable, among multiple factors, to be studied is bed grain size distribution. Grain size mapping and its monitoring overtime is necessary to understand geomorphological processes. Sediment dimensions and their pattern of change along rivers are important for mechanical and fluid dynamic properties of sediments, regulating channel morphology and hydraulics, as well as fluvial ecosystems functioning (Surian, 2001). Nowadays, we are aware of the importance of sampling and studying river catchments as holistic systems in order to support a successful river management. Without a large spatial and temporal understanding, river management cannot be fully effective and sustainable. In response to these requirements remotely sensed data are fundamental for hydromorphological assessment and monitoring, providing continuous data over entire catchments and allowing for a comprehensive river management. In the last two decades, Remote Sensing (RS) technology in river sciences has dramatically increased opening up unprecedented possibilities for river science and management. However, due to the high cost of RS data acquisition and the low spatial resolution available, field-based approaches are still dominating in river surveys and their prerequisites limit their application at the network scale. Recently, the most ambitious Earth observation programme to date, Copernicus, headed by the European Commission (EC) in partnership with the European Space Agency (ESA), has opened up unparalleled possibilities for environmental purposes and monitoring. ESA has developed a new family of missions called Sentinels that supply accurate, timely and easily accessible information. This innovation in RS satellite technology will provide an unparalleled amount of free data available to be investigated. The purpose of this work is to explore the potential of Sentinel 2 data on grain size mapping. Finding the relationship between multispectral information and grain size percentiles and assess the maximum attainable accuracy, creates the possibility to gain important information about broad classes of sediment along big rivers and their changes over time. Moreover, beside testing satellite potential, the broad scope is to find an efficient and repeatable methodology for grain size mapping. The solution we proposed, is the fusion of satellite and Unmanned Aerial Systems (sUAS) remote sensed imagery. The former provides the spectral information, the latter is useful to extract, by photosieving technique, ground truth grain size data, needed for validation. We tested this approach on the major Italian river, Po, in the north of Italy and we identified a functional model for describing the relationship between grain size percentiles and multispectral data. Results show that there is a significant relationship between satellite data and grain size percentiles and this represents a major advance in our current ability to characterize fluvial habitats and processes along major river systems, thanks to the characteristics of Sentinel 2 data, nowadays free available worldwide with a time frequency of about 5 days. The proposed methodology will allow for an objective and rapid grain size mapping, easily transferable to any large river areas, opening up broad perspectives for future fluvial survey practices.| File | Dimensione | Formato | |
|---|---|---|---|
|
2017_07_Marchetti.pdf
accessibile in internet per tutti
Descrizione: Testo della tesi
Dimensione
3.83 MB
Formato
Adobe PDF
|
3.83 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/135409