Hyperspectral Remote Sensing (RS) data have been exploited in the last decades to successfully retrieve information about plant properties at different scales. This thesis focuses on the use of state of the art hyperspectral RS data to retrieve vegetation status at canopy level, using both experimental and modeled data. In the last years, RS of Sun-induced chlorophyll fluorescence (F) emerged as a novel and promising scientific field for studying the dynamic behavior of photosynthesis. F is a physical side product of chlorophyll a light absorption that is emitted as an electromagnetic radiation in the red and far-red spectral regions (≈ 640 nm to 850 nm), and it is related to the energetic status of the photosystems. Nevertheless, apart from physiology F is concurrently influenced by leaf and canopy structure, pigment concentration and weather/illumination conditions, and its unambiguous interpretation is still challenging. This drives the interest in exploring F-derived metrics able to disentangle the physiological information from the remotely sensed F signal. In this thesis I analyzed data from two case studies: i) a managed loblolly pine forest (Parker Tract forest, North Carolina, U.S.A.) where I investigated how F and F-derived metrics vary with age-related changes in plant physiology; and ii) an experiment of induced stress where the photosynthesis of a homogeneous lawn was inhibited with a herbicide treatment. In the first case-study, I used hyperspectral data acquired with the HyPlant airborne sensor to characterize the F emission of 18 evenly aged stands in a range from 3 to 46 years old, and to calculate the Absorbed Photosynthetically Active Radiation (APAR). I computed the F yields in the red and far-red regions normalizing the corresponding F data for the APAR. Results show that in loblolly pine: i) red F and red F yield change with stand age, younger loblolly pines dissipate more red F than older one (up to 60% more) and the decline of F yield with stand age is more pronounced than that for red F; ii) the decline of red F yield can be related to the increase in water limitation occurring as loblolly trees grow in age and height. In the second case study three 9 m × 12 m plots of a homogeneous lawn were treated with different doses of a commercial formulation of Chlortoluron. This herbicide inhibits photosynthesis by blocking the electron transport chain in the photosynthetic apparatus. I exploited data collected on the ground with very high resolution spectroradiometers, airborne images collected with the HyPlant and the TASI-600 sensors, as well as canopy-level gas exchange measurements collected with closed chambers, in order to detect short-term dynamics of photosynthetic efficiency in vegetation, induced by stress. Results show that immediately after the application there was a rapid increase of F, F yields and the photosynthetic reflectance index (PRI). Canopy temperature also increased after the application of Dicuran. A later decrease of fluorescence and PRI was observed together with a reduction of chlorophyll a content and a drop of pigment-related vegetation indices. Moreover, the dosage of Chlortoluron had an impact in the dynamics of F. Starting from the ground level hyperspectral measurements over the lawn, I inverted numerically a simplified version of the SCOPE model to concurrently retrieve F, F yield and several biochemical and biophysical parameters of the vegetation from apparent reflectance data. For the first time the full spectrum of canopy F, the fluorescence yield, as well as the main vegetation parameters that control light absorption and reabsorption, were retrieved concurrently using canopy-level high resolution apparent reflectance measurements. The effects of pigment content, leaf/canopy structural properties and physiology were effectively discriminated. Their combined observation over time led to the recognition of dynamic patterns of stress adaptation and stress recovery.

Negli ultimi decenni il telerilevamento iperspettrale è stato utilizzato per stimare con successo informazioni sulle proprietà delle piante a diverse scale di indagine. Questa tesi si focalizza sull’uso di dati telerilevati iperspettrali per stimare lo stato di salute della vegetazione a livello di canopy. Recentemente, il telerilevamento passivo della fluorescenza della clorofilla indotta dal Sole (F) è emerso come un ambito scientifico di grande interesse per la studio del comportamento dinamico della fotosintesi. F è un prodotto dell’assorbimento della luce da parte della clorofilla a, emesso come radiazione elettromagnetica nel rosso e nel vicino infrarosso (≈ 640 nm to 850 nm), ed è relazionato allo stato energetico dei fotosistemi. Oltre che dalla fisiologia F è influenzata anche dalla struttura della foglia e della canopy, dalla concentrazione dei pigmenti e dalle condizioni di illuminazione/meteorologiche, e la sua interpretazione univoca è tuttora complessa. Questo guida l’interesse nell’esplorare metriche in grado di isolare l’informazione legata alla fisiologia presente nel dato di F. In questa tesi ho analizzato due casi studio: i) una foresta di pino loblolly (Parker Tract, North Carolina, U.S.A.) dove ho investigato la variazione di F e delle metriche derivate al variare dei processi fisiologici legati all’età delle piante; e ii) un esperimento di stress indotto dove è stata inibita la fotosintesi di un tappeto erboso. Nel primo caso di studio, utilizzato dati iperspettrali acquisiti con il sensore aereo HyPlant per caratterizzare la F emessa da 18 stand coetanei di pino di èta compresa tra 3 e 46 anni, e per calcolare la radiazione fotosinteticamente attiva assorbita (APAR). Ho calcolato gli yield di F normalizzando F per la APAR. I risultati mostrano che nel loblolly: i) gli yield di F nel rosso e nel vicino infrarosso cambiano al variare dell’età dello stand, i giovani pini dissipano più F nel rosso rispetto a quelli più maturi (fino al 60% in più) e il declino dello yield di F nel rosso con l’età dello stand è più pronunciato di quello di F; ii) il declino dello yield di F nel rosso può essere relazionato alla limitazione idraulica che si manifesta nel loblolly durante la crescita. Nel secondo caso di studio tre plot di 9m x 12m di un tappeto erboso sono stati trattati con diverse dosi di Chlortoluron. Questo erbicida inibisce la fotosintesi bloccando la catena di trasporto degli elettroni. Ho utilizzato dati acquisiti a terra con spettroradiometri ad alta risoluzione , immagini aeree con i sensori HyPlant e TASI-600, e misure di scambi gassosi, per studiare le dinamiche a breve termine dell’efficienza fotosintetica della vegetazione, indotte dallo stress. I risultati mostrano che subito dopo il trattamento si è verificato un rapido incremento di F, dello yield di F, della temperatura della canopy e del Photosynthetic Reflectance Index (PRI). Successivamente è stato osservato un decremento dose-specifico di F e del PRI, assieme a una riduzione del contenuto di clorofilla a e di altri indici di vegetazione legati ai pigmenti. A partire dai dati spettrali acquisiti a terra sul tappeto erboso, ho invertito numericamente una versione semplificata del modello SCOPE, per stimare per la prima contemporaneamente lo spettro completo di F, lo yield di F, assieme ai principali parametri della vegetazione che controllano l’assorbimento e il riassorbimento della luce. L’effetto del contenuto di pigmenti, delle proprietà strutturali della foglia e della canopy, e la fisiologia sono state discriminate con successo. La loro osservazione combinata nel tempo ha portato al riconoscimento di pattern dinamici di risposta e di adattamento allo stress.

(2018). Development of novel methods to evaluate vegetation status from multi-source remote sensing data. (Tesi di dottorato, Università degli Studi di Milano-Bicocca, 2018).

Development of novel methods to evaluate vegetation status from multi-source remote sensing data

CELESTI, MARCO
2018

Abstract

Hyperspectral Remote Sensing (RS) data have been exploited in the last decades to successfully retrieve information about plant properties at different scales. This thesis focuses on the use of state of the art hyperspectral RS data to retrieve vegetation status at canopy level, using both experimental and modeled data. In the last years, RS of Sun-induced chlorophyll fluorescence (F) emerged as a novel and promising scientific field for studying the dynamic behavior of photosynthesis. F is a physical side product of chlorophyll a light absorption that is emitted as an electromagnetic radiation in the red and far-red spectral regions (≈ 640 nm to 850 nm), and it is related to the energetic status of the photosystems. Nevertheless, apart from physiology F is concurrently influenced by leaf and canopy structure, pigment concentration and weather/illumination conditions, and its unambiguous interpretation is still challenging. This drives the interest in exploring F-derived metrics able to disentangle the physiological information from the remotely sensed F signal. In this thesis I analyzed data from two case studies: i) a managed loblolly pine forest (Parker Tract forest, North Carolina, U.S.A.) where I investigated how F and F-derived metrics vary with age-related changes in plant physiology; and ii) an experiment of induced stress where the photosynthesis of a homogeneous lawn was inhibited with a herbicide treatment. In the first case-study, I used hyperspectral data acquired with the HyPlant airborne sensor to characterize the F emission of 18 evenly aged stands in a range from 3 to 46 years old, and to calculate the Absorbed Photosynthetically Active Radiation (APAR). I computed the F yields in the red and far-red regions normalizing the corresponding F data for the APAR. Results show that in loblolly pine: i) red F and red F yield change with stand age, younger loblolly pines dissipate more red F than older one (up to 60% more) and the decline of F yield with stand age is more pronounced than that for red F; ii) the decline of red F yield can be related to the increase in water limitation occurring as loblolly trees grow in age and height. In the second case study three 9 m × 12 m plots of a homogeneous lawn were treated with different doses of a commercial formulation of Chlortoluron. This herbicide inhibits photosynthesis by blocking the electron transport chain in the photosynthetic apparatus. I exploited data collected on the ground with very high resolution spectroradiometers, airborne images collected with the HyPlant and the TASI-600 sensors, as well as canopy-level gas exchange measurements collected with closed chambers, in order to detect short-term dynamics of photosynthetic efficiency in vegetation, induced by stress. Results show that immediately after the application there was a rapid increase of F, F yields and the photosynthetic reflectance index (PRI). Canopy temperature also increased after the application of Dicuran. A later decrease of fluorescence and PRI was observed together with a reduction of chlorophyll a content and a drop of pigment-related vegetation indices. Moreover, the dosage of Chlortoluron had an impact in the dynamics of F. Starting from the ground level hyperspectral measurements over the lawn, I inverted numerically a simplified version of the SCOPE model to concurrently retrieve F, F yield and several biochemical and biophysical parameters of the vegetation from apparent reflectance data. For the first time the full spectrum of canopy F, the fluorescence yield, as well as the main vegetation parameters that control light absorption and reabsorption, were retrieved concurrently using canopy-level high resolution apparent reflectance measurements. The effects of pigment content, leaf/canopy structural properties and physiology were effectively discriminated. Their combined observation over time led to the recognition of dynamic patterns of stress adaptation and stress recovery.
ROSSINI, MICOL
COLOMBO, ROBERTO
remote; sensing;; hyperspectral;; fluorescence;; vegetation
remote; sensing;; hyperspectral;; fluorescence;; vegetation
GEO/10 - GEOFISICA DELLA TERRA SOLIDA
English
5-mar-2018
SCIENZE CHIMICHE, GEOLOGICHE E AMBIENTALI - 94R
30
2016/2017
open
(2018). Development of novel methods to evaluate vegetation status from multi-source remote sensing data. (Tesi di dottorato, Università degli Studi di Milano-Bicocca, 2018).
File in questo prodotto:
File Dimensione Formato  
phd_unimib_758963.pdf

accesso aperto

Descrizione: tesi di dottorato
Tipologia di allegato: Doctoral thesis
Dimensione 13.95 MB
Formato Adobe PDF
13.95 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/10281/199119
Citazioni
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
Social impact