Im Kolloquium wird das wissenschaftliche Arbeiten reflektiert und das Management des Projektes Bachelorarbeit diskutiert. Dies beinhaltet die begründete Auswahl eines Arbeitsthemas mit der Einbindung in eine gesellschaftsrelevante Problemstellung und einen theoretischen Kontext. Dazu gehört die Entwicklung konkreter Zielsetzungen und wissenschaftlicher Fragestellungen sowie die Bildung von Hypothesen. Hierauf basierend sind das methodische Vorgehen und ein Arbeitsplan auszuarbeiten und zu begründen. Abschließend sind die Möglichkeiten der Ergebnisinterpretation und ihr Rückbezug zu Fragestellungen, Zielsetzungen und Hypothesen zu diskutieren.

Der Betreuer bzw. die Betreuerin der Bachelorarbeit ist frei wählbar, die Betreuung der Arbeit ist nicht an den Seminarleiter gebunden. Es sollen möglichst konkrete Bachelorarbeiten vorgestellt und besprochen werden. Notfalls kann auch ein fiktives Thema gewählt werden.

Das Seminar findet alle zwei Wcohen zweistündig stat. Zu Beginn des Semesters mit der Theorie und im Laufe des Semesers dann mit freier Arbeitspahse zur Erstellung des Exposeekonzepts und Vorstellung der Exposees. Abgabe der schriftliche Fassung des Exposees: 26.01.24.  

The state-of-the-art remote sensing of vegetation uses data collected with various optical sensors on towers, drones, aircraft, and satellite platforms. The recent boom of data-driven machine learning algorithms (e.g., deep-learning neural networks, random forests, or Gaussian processes) allows for an efficient interpretation of such multiscale and multiresolution data, transforming the optical signals into specific information about vegetation functional traits. Yet, to be successful, this approach requires a robust and comprehensive knowledge base for i) achieving a proper training of interpretational algorithms, ii) understanding what type of optical inputs are most suitable, and iii) optimizing sensor and data technical specifications. Coupled leaf and canopy physical radiative transfer models (RTMs), simulating interactions of electromagnetic radiation within plant canopies, provide virtual environments suitable to generate such required knowledgebase.

In this course, you will learn how to map quantitative plant functional traits, such as leaf chlorophyll content, leaf water content, leaf area index, etc., from spectral satellite data (e.g., Sentinel-2) using a combination of the physical RTMs and modern machine-learning methods. Upon successful completion of this module, you will know how to:

work and pre-process the satellite spectral images of terrestrial vegetation formations,
parameterize and run in a forward mode leaf and canopy RTMs in order to simulate virtual spectral satellite observations,
train machine-learning models properly using RTMs simulated remote sensing data of the specific vegetation types (e.g., crops),
apply the machine-learning methods to quantify plant functional traits from the satellite observations, and
interpret the obtained maps, including the related uncertainty estimates.
The seminar will start with a 20-30 min long theoretical lecture related to the topic of this course. It will be followed up by a practical 'hands-on' part, where you will use the ARTMO toolbox to conduct sequentially retrieval of specific plant traits from satellite optical imagery.

Communication language (spoken as well as written): English

In 2016 and 2019, Australia, especially the Tasmanian island, was subjected to some of the most widespread and devastating wildfires in recent history. What made these fires so devastating is the particularly ”dense bush” type of Australian vegetation that they affected. In 2016, large areas within the Tasmania World Heritage Area and the Arthur-Pieman Conservation area were destroyed. These vegetation communities are unique and take thousands of years to regrow. Many of the affected areas are remote and largely inaccessible, making a ground assessment of the true scale of the fire devastation difficult, in some locations even impossible. In this Geomatic Method seminar, you will use multispectral remote sensing images of Sentinel-2A and Landsat-8 satellites to identify the extent of the wildfires within the Arthur-Pieman Conservation area and estimate the types and amounts of vegetation that were affected by the 2016 burn.

Participants will individually execute 12 assignment tasks. The tasks will be split into two parts:

literature work (2 tasks), resulting in a short (400 words max.) essay describing theoretical bases of forest fire mapping from multispectral space-borne image data (written preferably in English), and
computer work, comprised of 10 practical tasks and several questions about satellite image pre-processing methods (e.g., spatial and radiometric corrections), spectral analyses (e.g., automatic image classifications), and landcover change detection (e.g., basic statistics). The results will be submitted to the lecturer as a report, i.e., *.PDF or *.DOCX digital document written in English.
Each of the 2 x 45 min seminars will consist of:

15-minute theoretical introduction related to the assigned task and presented by the lecturer,
5-10-minutes of discussion about general issues and problems related to individual tasks, and
about 65 min of individual hands-on computer work under the supervision of the lecturer.
The final mark will reflect the quality of the literature work delivered in form of an essay (30% of the overall mark) and the correctness of the satellite image analyses’ outputs and answers to questions given in the computer work part (70% of the overall mark).

On successful completion of this module, you will know how to:

search for publicly available satellite imagery in online repositories,
use scientific literature to search for and process relevant published information,
apply satellite image processing techniques in the context of forest fire detection and bunt area assessment, and
extract quantitative and thematic information from satellite multispectral imagery by means of image pre-processing, classification, and change detection techniques.
The professional remote sensing software ENVI will be used in this module for computer-based work. The basic functionality of this software, i.e., loading and displaying an image with different spectral band combinations, image composition contrast enhancement, plotting of spectral signatures, work with regions of interest (ROIs), application of required image analyses, and export of images and plots suitable for inclusion in your assignment report, will be explained during first seminar sessions. Previous knowledge about remote sensing image processing will be to your advantage.

If you have questions or need further information please contact rsrgedu@uni-bonn.de.

Course motivation and content:

In the Anthropocene era, impacted by a man-caused global warming due to excessive CO2 emissions, computer models simulating and forecasting temporally dynamic eco-physiological processes of terrestrial vegetation, such as carbon uptake or evapotranspiration, are essential for a ‘carbon neutral’ economy and a ‘carbon reducing’ management of natural ecosystems. Optical remote sensing provides multitude of input data for dynamic land surface vegetation modelling, spanning from categorical maps of land cover classes to spatially explicit quantitative estimates of vegetation biochemical and physical traits such as: fraction of absorbed photosynthetic radiation, leaf area index, or contents of water, photosynthetic foliar pigments, and nitrogen. This remotely sensed information, retrieved from air-/space-borne spectral and thermal sensors, is getting more accurate and more frequent thanks to the improving sensing technology (e.g., hyperspectral imaging spectroscopy), a higher number of satellite constellations (e.g., new Sentinel and future Chime ESA satellites), and newly available optical signals like, for instance, sun-induced chlorophyll fluorescence (retrievable from NASA’s OCO-2/3 and ESA’s TROPOMI and future FLEX missions). This course is designed to enlarge knowledge and deepen understanding of M.Sc. students about actual and near future optical vegetation remote sensing products that can be assimilated by existing eco-physiologically based land surface vegetation models, such as ORCHIDEE, JULES or LPJ-GUESS. The knowledge gained in this course is essential for future study climate change related topics and useful for work in fields of precision agriculture, forest monitoring and natural ecosystems’ management.

Course structure:

To comprehend advanced approaches for quantitative estimations of vegetation traits from Earth observations, the course participants will learn how the solar electromagnetic radiation interacts with individual plant leaves and canopies, what are the main pre-processing and remote sensing image interpretation methods, and how are the obtain maps of vegetation traits assimilated by the land surface vegetation models. 

The seminar will be split into three thematic parts:

Remote sensing observations: Acquisition, calibration, and pre-processing of leaf and canopy spectral measurements data.
Interpretation methods: Methods for retrieval of vegetation traits from satellite spectral measurements.
Dynamic vegetation models: Assimilation of remote sensing products in selected land surface vegetation models.
The seminar will contain theoretical lectures (45-min per week) delivered by the lecturer and computer-based practical exercises (for instance, spectroscopy of vegetation leaves and canopies) performed by the course participants. A multiple-choice test will be carried out upon the completion of each thematic part. The final grade of each seminar participant will be deduced from correct answers of tests and computer exercises.

Course outcomes:

On successful completion of this module, students will be able to:

Understand how optical remote sensing can inform models predicting vegetation functioning within the carbon and water cycles.
Find and assess quality of remotely sensed maps of vegetation traits relevant for eco-physiological modelling.
Use/assimilate vegetation remote sensing product in to the dynamic land surface vegetation models in order to predict, for instance, vegetation carbon assimilation capacity or water use efficiency.  
 

Language of communication (both oral and written) is English.

Course motivation

The ongoing global climate change, mainly the increase in air temperature caused by human activities since the industrial revolution, is impacting not only natural processes and cycles but also human societies and their economic activities on a global scale. We are witnessing, for instance, unprecedently severe forest fires in the northern America, la Niña induced floods in eastern Australia, extensive thawing of permafrost in Siberia and Alaska, unexpected drying of large water bodies due to the freshwater shortage, and rapid melting of glaciers in the Arctic regions and high mountains. In this project seminar module, we will use optical remote sensing image data acquired from satellites orbiting the Earth (e.g., Copernicus/ESA’s Sentinel-2 and NASA’s Landsat-8 & 9) or from hyperspectral visible, near-infrared, and thermal images from aircrafts (e.g., airplanes) to detect, map, and potentially monitor in space and time the actual impacts of climate change specific, but not exclusive, events on the Earth surface.

Course structure

Participants will form small research teams (groups of approx. 3 students) and work together on one of the climate change-related topics of their own liking or provided by the lecturer. Examples of project topics are:

Drying of the Aral Sea and its consequent impact on local activities/land use,
Outbreaks of insect parasites, e.g., bark beetle, in mid European temperate forests,
Contribution of thawing Arctic permafrost to regional land cover changes,
Mapping impacts of Monsoon-induced floods at Asian and Australian continents,
Assessing natural and economic impacts of big forest fires in California in 2020,
Cities as summer heat islands – existing causes and potential remediations.
Weekly 4-hour courses will include:

30-45-minute theoretical lecture introducing the state-of-the-art remote sensing techniques relevant to your projects,
short presentation by a member of each research team (each week different team member) outlining the status and progress of the project work, and
actual teamwork with hands-on project work supported by the lecturer.
Course outcomes

Upon successful completion of this module, you will know how to:

find online available optical remote sensing satellite data and process it in the state-of-the-art remote sensing software,
design and execute a research project, including a proper definition of scientific hypothesis/questions, selection of appropriate optical remote sensing input data and methods, and make synthesis of analytical results answering the research questions,
write a report in accordance with the scientific standards, including proper presentation of graphical outputs (e.g., maps) and correct referencing of the relevant scientific literature (research papers, books, and conference proceedings), and
work in a team, present your intentions and results in a concise and understandable manner, receive feedback from your peers and disseminate it to improve your work and conduct an evidence-based scientific discussion.

Course motivation and content:

With the increasing availability of satellite imagery, time series of remotely sensed datasets are now widely used to monitor various aspects of the Earth environment, including land surfaces, open water and atmosphere, and their interactions with human activities under the on-going global climate change.

Recently, Google Earth Engine (GEE) has become one of the most widely used tools in processing long-term satellite observations time series for scientific as well as policy-making purposes. GEE provides free access to large-scale data analysis using an online ‘cloud’ computing technology. Its Java language-based interface allows users to access and analyze satellite image data without the need for local storage or software installation.

This seminar will focus on time series analysis using the GEE Code Editor, by combining theoretical lectures with practical ‘hands-on’ programming exercises.

Course structure:

Participants will work individually or in pairs on weekly topics related to their specific areas of interest. These topics will focus on the analysis of time series data to monitor, detect, or provide insight into the following topics:

-        Variation of water surface areas to assess the impact of drought and flood events and to analyze the frequency of these events over a given period.

-        Temporal changes of different land cover types, including agricultural areas, evergreen forests, urban areas, and water bodies.

-        Changes in forest areas assessing deforestation of the evergreen forest region.

-        Change of agricultural area due to factors such as limited water supply, heat waves, and other relevant environmental impacts.

-        Monitoring of urban area to observe urban expansion over time.

-        Seasonal changes in an evergreen forest area, based on vegetation spectral indices, to investigate a relationship with an increasing pressure of the climate change.

-        The impact of coal mining on productive agricultural landscapes.

-        Dense time series analysis using the Breaks For Additive Season & Trend (BFAST) software toolbox.

Weekly 2-hour courses will include:

-      30-45 minute theoretical lecture introducing the state of the art of remote sensing time series data and analysis methods.

-      After the lecture, students will be provided with a step-by-step (function-by-function) guide to programming in GEE, including sample JavaScripts to perform pre-defined GEE analyses.   

Course outcomes:

On successful completion of this module, students will be able to:

-      Select appropriate time series remote sensing input data, pre-processing method, and analysis methods.

-      Pre-process satellite data timeseries by mastering smoothing, filtering, and filling missing data techniques.

-      Programming with JavaScript in GEE.

-      Carry out remote sensing analysis quality assessment, such as, collecting reference data and cross-validation to assess the accuracy and reliability of the analysis.     

Language of communication (both oral and written) is English.

Software used in this course: https://earthengine.google.com/#!/ Specific study literature will be recommended by lecturer during the seminar lectures

In diesem Seminar werden verschiedene Aspekte zur Nutzung von Cloud Plattformen, deren Datenangebot und Schnittstellen vorgestellt. Ein Focus wird hierbei auf Applikationen zum Einsatz von Satellitendaten, insbesondere von Daten des Copernicus Programms der Europäischen Raumfahrtagentur (ESA), liegen. Es wird ebenso die Nutzung der Daten der Copernicus Kerndienste (Landüberwachung, Meeresumwelt, Atmosphäre, Klimawandel, Katastrophen- und Krisenmanagement und Sicherheit) vorgestellt.

Die nationalen Plattformen CODE-DE und EO-Lab, wie auch das europäische Copernicus DataSpace Ecosystem (CDSE) werden für Applikationen genutzt, wie auch die Google Earth Engine.

Das Seminar baut auf einem ersten theoretischen Teil auf, bei dem auch die für das Seminar relevanten Grundlagen der Fernerkundung wiederholt werden. Im einem zweiten Teil sollen existieren Daten-Dienste vorgestellt werden, sowie in einem dritten Teil der Einstieg in das Cloud basierte Arbeiten mit Satellitendaten erfolgen.

Im wissenschaftlichen Kolloquium zur Masterarbeit werden laufende Masterarbeiten und Forschungsarbeiten am Zentrum für Fernerkundung der Landoberfläche (ZFL) https://www.zfl.uni-bonn.de/ und der Arbeitsgruppe Fernerkundung (RSRG) am Geographischen Institut vorgestellt und diskutiert. Begleitend setzen wir uns mit aktuellen wissenschaftlichen Aufsätzen und Methoden der Erdbeobachtung und Themen des Raumfahrtmanagements auseinander. Anhand der Master-/Forschungsarbeiten und Zeitschriftenartikel werden die vielfältigen Aspekte des wissenschaftlichen Arbeitens erarbeitet (Fragestellung, Hypothesenbildung, Aufbau/Struktur, etc.).

Es ist geplant, dass regelmäßig externe Wissenschaftler*innen aus Nachbardisziplinen und von anderen Universitäten und Instituten zu Gast im Kolloquium sein werden. Anhand der Vorträge werden verschiedene aktuelle Forschungsthemen, wissenschaftliche Fragestellungen und Herangehensweisen aus der Erdbeobachtung veranschaulicht.

Von den Teilnehmer*innen der Lehrveranstaltung wir eine aktive und regelmäßige Teilnahme, die Ausarbeitung eines Exposées für die Masterarbeit sowie die Präsentation der laufenden Masterarbeit oder einer anderen Forschungsarbeit erwartet.

Das Seminar findet alle zwei Wochen zweistündig statt.

Level: Master

Kurs: M4 Projektseminar

Credits: 18LP - 6SWS

The ongoing climate change, namely the increase in air temperature caused by human activities since the industrial revolution, is impacting not only natural processes and cycles but also human societies and their economics on a global scale. We are witnessing, for instance, unprecedently severe forest fires in the northern hemisphere and Australia, extensive thawing of permafrost, unexpected drying of large water bodies due to the freshwater shortage, rapid melting of glaciers in the Arctic regions and high mountains. In this project seminar module, we will use optical remote sensing image data acquired from satellites (e.g., Copernicus/ESA’s Sentinel-2 and NASA’s Landsat-8) or aircraft (hyperspectral visible, near-infrared, and thermal images) to detect and map the actual impacts of some of the climate change events in space and time.

Upon successful completion of this module, you will know how to:

• design and execute a research prokject, including a proper definition of scientific hypothesis/questions, selection of appropriate optical remote sensing input data and processing methods, and make a synthesis of analytical results answering the research questions 
• work efficiently with optical remote sensing image data and state-of-the-art image processing software,
• write a scientific report in accordance with current standarts, inclduing proper presentation of graphical outputs (e.g. maps) and correct referencing of the relevant scientific communications, and
• work in a team, present your intentions and results in a concise and understandable manner, recieve feedback from your peers and disseminate it to improve your work, and conduct an evidence-based scienitifc discussion.

The remote sensing image analyses and interpretaions will be done in a professional remote sensing software ENVI (QGIS  might be considered). A basic understanding of remote sensing image processing is required.

Level: Master

Kurs: M2 Forschungsmethoden

Credits: 6LP - 2SWS

In 2016 and 2019, Australia, especially Tasmania, was subjected to some of the most widespread and devastating forest fires in recent history. What made these fires so devastating is the particularly “dense bush” type of Australian vegetation they affected. In 2016, large areas within the Tasmania World Heritage Area and the Arthur-Pieman Conservation area were destroyed. These vegetation communities are unique and take thousands of years to regrow. Many of the affected areas are remote and largely inaccessible, making proper assessment of the true scale of the fire devastation difficult. For this Geomatic Method seminar, you will use multispectral remote sensing images of Sentinel-2A and Landsat-8 satellites to identify the extent of the blaze within the Arthur-Pieman Conservation area and estimate the types and amounts of vegetation affected by the 2016 burn.

On successful completion of this module, you will know how to:

• search for publicly available satellite imagery in the online repositories,
• use scientific literature to search for and process relevant published information,
• apply satellite image processing techniques in the context of forest fire detection and bunt area assessment, and
• extract quantitative and thematic information from satellite multispectral imagery by means of image pre-processing, classification, and change detection techniques.

In this module, the professional remote sensing software ENVI will be used for computer-based work. The basic functionality of this software, i.e., loading and displaying an image with different spectral band combinations, image composition contrast enhancement, plotting of spectral signatures, work with regions of interest (ROIs), application of required image analyses, and export of images and plots suitable for inclusion in your assignment report, will be explained (if needed) during this module. Previous knowledge about remote sensing image processing will be to your advantage.

Level: Bachelor

Kurs: B8 I/II Forschungsmethoden

Credits: 6LP - 2SWS

Remote sensing technologies and methodologies are advancing at a fast pace, getting more physically based, data-driven and sophisticated. This course is designed as a continuation of Introduction to Remote Sensing Mapping and Environmental monitoring (B7). Participants will learn how to apply some of the concepts they have already studied and incorporate the actual remote sensing spectral and radar techniques to solve societal challenges. They will apply advanced classification and regression methods, such as machine learning, and various quantitative retrievals from hyperspectral image data, to solve some of the current environmental and scientific problems.

Level: Bachelor

Kurs: B11 Projektseminar

Credits: 12LP - 4SWS

The ongoing global climate change, mainly the increase in air temperature caused by human activities since the industrial revolution, is impacting not only natural processes and cycles but also human societies and their economic activities on a global scale. We are witnessing, for instance, unprecedently severe forest fires in the northern America, la Niña induced floods in eastern Australia, extensive thawing of permafrost in Siberia and Alaska, unexpected drying of large water bodies due to the freshwater shortage, and rapid melting of glaciers in the Arctic regions and high mountains. In this project seminar module, we will use optical remote sensing image data acquired from satellites orbiting the Earth (e.g., Copernicus/ESA’s Sentinel-2 and NASA’s Landsat-8 & 9) or from hyperspectral visible, near-infrared, and thermal images from aircrafts (e.g., airplanes) to detect, map, and potentially monitor in space and time the actual impacts of climate change specific, but not exclusive, events on the Earth surface.

Schwerpunkte im Teilbereich Fernerkundung:

• Theoretische und physikalische Grundlagen der FE
  Aufnahmesysteme: Sensoren, Plattformen, Bildformate
• Aufbereitung und inhaltliche
  Auswertung von Satellitendaten
• Fallbeispiele für geographische Anwendungen (z.B. Landnutzungsänderungen,
  Städtewachstum, Landdegradation, Gletscherschmelze)

Qualifikationsziele im Teilbereich Fernerkundung:

• Kenntnisse in die physikalischen Grundlagen der Fernerkundung (FE)
• Kenntnisse in der visuellen Bildinterpretation und in der
• Aufbereitung von
  digitalen Satellitendaten
• Kenntnisse in der inhaltlichen Auswertung von FE-daten (z.B. Bildklassifikation,
  Veränderungsdetektion, Zeitreihenanalyse)

In this course, you will learn how to map quantitative plant functional traits, such as leaf chlorophyll content, leaf water content, leaf area index, etc., from spectral satellite data (e.g., Sentinel-2) using a combination of the physical RTMs and modern machine-learning methods. Upon successful completion of this module, you will know how to:

• work and pre-process the satellite spectral images of terrestrial vegetation formations

• parameterize and run in a forward mode leaf and canopy RTMs in order to simulate virtual spectral satellite observations,

• train machine-learning models properly using RTMs simulated remote sensing data of the specific vegetation types (e.g., crops),

• apply the machine-learning methods to quantify plant functional traits from the satellite observations, and

• interpret the obtained maps, including the related uncertainty estimates.

Growing human population density combined with the ongoing climate change, both taking place on a global scale, are increasing pressure on the Earth's resources and services provided by ecosystems all over the World. Appropriate management and adequate solutions to environmental problems, such as deforestation, overgrazing, soil contamination and depletion, and water shortage, are heavily dependent on accurate and timely knowledge. To work collaboratively with experts from multiple disciplines, planners, managers, policymakers, and researchers require a comprehensive understanding of the complex factors involved in processes driving the environmental problems and challenges. Here, remote sensing data and their interpretation play a central role in the quest for required knowledge.

Participants will form small research teams (groups of approx. 3 students) and work together on one of the climate change-related topics of their own liking or provided by the lecturer. The module will be split into two parts: I) a block of morning and afternoon intensive sessions (4 + 4 SWS between 9.00 and 17.00 hrs) from 10 until 14 October 2022, and II) a morning session each second week from 20 October 2022 till 02 February 2023. 

Upon successful completion of this module, you will know how to:

• design and execute a research project, including a proper definition of scientific hypothesis/questions, selection of appropriate optical remote sensing input data and processing methods, and make a synthesis of analytical results answering the research questions,
• work efficiently with optical remote sensing image data and the state-of-the-art image processing software,
• write a scientific report in accordance with the current standards, including proper presentation of graphical outputs (e.g., maps) and correct referencing of the relevant scientific communications (research papers, books, and conference proceedings), and
• work in a team, present your intentions and results in a concise and understandable manner, receive feedback from your peers and disseminate it to improve your work and conduct an evidence-based scientific discussion.