Guillermo CastillaNatural Resources Canada | NRCan · Canadian Forest Service
Guillermo Castilla
PhD
About
60
Publications
56,219
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
3,059
Citations
Introduction
Guillermo is a remote sensing research scientist who collaboratively develops mapping and monitoring methods, delivers geospatial information, and generates and disseminates knowledge relevant to Canada’s forests, from individual trees to entire ecozones.
Additional affiliations
February 2014 - present
May 2006 - January 2014
July 2010 - May 2017
Publications
Publications (60)
Background Satellite imaging has improved burned area mapping; however, few studies have taken advantage of the Multi-Spectral Scanner (MSS) in early Landsat satellites, which started acquiring data 10 years earlier than Thematic Mapper (TM). Aims To expand Canada’s National Burned Area Composite (NBAC) annual time series back to 1972 using MSS dat...
Black spruce is the most common tree species in Canada, dominating treed peatlands where they are usually stunted. We used 495 destructively sampled trees from 56 plots to develop allometric models of aboveground biomass by components (stem, branches, and needles) for small (<5 m tall) black spruce from peatlands in the Taiga Plains and Boreal Plai...
Operational forest monitoring often requires fine-detail information in the form of an orthomosaic, created by stitching overlapping nadir images captured by aerial platforms such as drones. RGB drone sensors are commonly used for low-cost, high-resolution imaging that is conducive to effective orthomosaicking, but only capture visible light. Therm...
Wildfires are a major natural disturbance in Canada that are postulated to increase under a warming climate. To derive accurate trends in burned area and to quantify the effects of fire frequency, duration, and extent, a sufficiently long time series of reliable burned area maps is required. With that in mind, we extended Canada’s National Burned A...
Satellite forest inventories are the only feasible way to map Canada’s vast, remote forest regions, such as those in the Northwest Territories (NWT). A method used to create such inventories is the k-nearest neighbour (k-NN) algorithm, which spatially extends information from forest inventory (FI) plots to the entire forest land base using wall-to-...
Sustainable forest management requires information on the spatial distribution, composition, and structure of forests. However, jurisdictions with large tracts of noncommercial forest, such as the Northwest Territories (NWT) of Canada, often lack detailed forest information across their land base. The goal of the Multisource Vegetation Inventory (M...
Research Highlights: This study advances the effort to accurately estimate the biomass of trees in peatlands, which cover 13% of Canada’s land surface. Background and Objectives: Trees remove carbon from the atmosphere and store it as biomass. Terrestrial laser scanning (TLS) has become a useful tool for modelling forest structure and estimating th...
Research Highlights: This is the most comprehensive analysis to date of the accuracy of height estimates for individual conifer seedlings derived from drone-based image point clouds (DIPCs). We provide insights into the effects on accuracy of ground sampling distance (GSD), phenology, ground determination method, seedling size, and more. Background...
There is uncertainty over the nature of ecosystem processes and trajectories that underlie NDVI 'greening’ and ‘browning' trends in sub-Arctic environments. In extreme cases, such as landcover change due to wildfire, there is justification for equating 'greening' (increasing NDVI) to an increase in biomass, and 'browning' (decreasing NDVI) to a los...
There is uncertainty over the nature of ecosystem processes and trajectories that underlie NDVI 'greening' and 'browning' trends in sub-Arctic environments. In extreme cases, such as landcover change due to wildfire, there is justification for equating 'greening' (increasing NDVI) to an increase in biomass, and 'browning' (decreasing NDVI) to a los...
Monitoring tree regeneration in forest areas disturbed by resource extraction is a requirement for sustainably managing the boreal forest of Alberta, Canada. Small remotely piloted aircraft systems (sRPAS, a.k.a. drones) have the potential to decrease the cost of field surveys drastically, but produce large quantities of data that will require spec...
Coarse woody debris (CWD; large parts of dead trees) is a vital element of forest ecosystems, playing an important role in nutrient cycling, carbon storage, fire fuel, microhabitats, and overall forest structure. However, there is a lack of effective tools for identifying and mapping both standing (snags) and downed (logs) CWD in complex natural se...
Rapid assessment of forest regeneration using unmanned aerial vehicles (UAVs) is likely to decrease the cost of establishment surveys in a variety of resource industries. This research tests the feasibility of using UAVs to rapidly identify coniferous seedlings in replanted forest-harvest areas in Alberta, Canada. In developing our protocols, we ga...
The oil and gas industry has grown significantly throughout the boreal and arctic ecosystems of North America. A major feature of the ecological footprint of oil and gas exploration is seismic lines—narrow corridors used to transport and deploy geophysical survey equipment. These lines, which traverse forests, tundra, uplands, and peatlands, were h...
![Vegetation height measurement at a sample point (modified from [34]).](publication/321505044/figure/fig4/AS:11431281211955916@1702512342536/Vegetation-height-measurement-at-a-sample-point-modified-from-34_Q320.jpg)
Monitoring vegetation recovery typically requires ground measurements of vegetation height, which is labor-intensive and time-consuming. Recently, unmanned aerial vehicles (UAVs) have shown great promise for characterizing vegetation in a cost-efficient way, but the literature on specific methods and cost savings is scant. In this study, we surveye...
Wildfires are burning increasingly extensive areas of forest in Canada, reducing their
capacity as carbon sinks. Here we compare the answers that four independent land cover datasets,
produced from different satellite images (SPOT, Landsat, and MODIS), provide for the question
of what burned in Canada in recent years. We harmonized the different da...
Outbreaks of insect pests periodically cause large losses of volume in Canada’s forests. Compounded with climate change, outbreaks create significant challenges for managing the sustainable delivery of ecosystem services. Current methods to monitor damage by these pests involve both field and aerial surveys. While relatively cost effective and time...
Despite recent calls for statistically robust and transparent accuracy assessment [1], further attention to rigor is still needed. Here I take the opportunity of a disputed accuracy assessment recently published in Remote Sensing [2,3,4] to highlight some issues regarding sampling design, response design and analysis that I often find as a reviewer...
The analysis of multitemporal sequences of remote sensing imagery is becoming increasingly important as Earth Observation archives continue to grow. While new methods and algorithms are continuously being developed, both for Landsat and MODIS, work in the change-visualization front is scant. In particular, new methods are needed that provide a mean...
We present the Change, Aftereffect, and Trend (CAT) transform for visualizing and analyzing landscape dynamics from dense, multi-annual satellite vegetation index (VI) time series. The transform compresses a temporally detailed, multi-annual VI dataset into three new variables capturing change events and trends occurring within that period. First,...
We describe the creation of a GIS vector layer of land cover polygons for the entire province of Alberta, Canada, based upon preexisting, Landsat-derived, land cover raster datasets circa 2000 that were produced by the Canadian federal government. Our novel spatial and semantic generalization algorithm begins with a morphological segmentation of th...
We introduce POLS (shorthand for POLygon Sampling), a versatile GIS tool for extracting a random subset of polygons from a vector layer. POLS enables users to optionally (i) set the sampling intensity in terms of percent area of the layer, number of polygons, or both; (ii) specify different strata to be sampled with equal or different intensity; (i...
We recently completed the accuracy assessment of a Landsat-derived landcover polygon layer covering the entire province of Alberta (660,000 km2), Canada, for which we gathered reference information for nearly 5000 randomly selected polygons ranging from two hectares to thousands of hectares in size. This gave us the unique opportunity to quantify,...
Conventional forests inventories narrowly focus on timber attributes and often neglect other aspects that may be relevant for other purposes. In an effort to broaden the usefulness of these inventories, we introduce a new protocol based on softcopy photo-interpretation for efficiently capturing both natural and anthropogenic features across a varie...
Vector-based landcover (LC) maps derived from GEographic Object-Based Image Analysis (GEOBIA) are increasingly replacing the
traditional raster maps from per-pixel classification, but our strategies for assessing their quality are not yet fully developed. We
contend that a complete accuracy assessment of a vector LC map must provide answers to the...
Modern ecological models often account for the influence of the surrounding environment by using landscape pattern indices (LPIs) as measures of landscape structure. Ideally, the landscape samples from which these LPIs are extracted should be centered on the locations where the response variable was measured. However, in situations where this is no...
Monitoring landscape change is a requisite for sustainable development that should be achievable through the analysis of multitemporal satellite imagery. However, the development of effective methods to analyze these data in a consistent and reliable way is still a challenging issue that demands new approaches. Here we demonstrate the use of a rece...
Unfortunately, many GIS layers depicting transportation networks do not provide information on the construction year of each line segment in the network. This poses a serious problem when the GIS layer is used as input to historic analyses investigating the growth of the human footprint in an area still being developed, since there is no way of fin...
Lidar (light detection and ranging) has demonstrated the ability to provide highly accurate information on forest vertical structure; however, lidar data collection and processing are still expensive. Very high spatial resolution optical remotely sensed data have also shown promising results to delineate various forest biophysical properties. In th...
We argue that thematic resolution, i.e., the level of categorical detail of a thematic map expressed by the number of classes
included in the map legend, is an inherent component of the scale at which a landscape is analyzed. Changing the number of
classes can change the configuration of the patch mosaic as much as changing the grain does. We addre...
LiDAR canopy height models (CHMs) can exhibit unnatural looking holes or pits, i.e., pixels with a much lower digital number than their immediate neighbors. These artifacts may be caused by a combination of factors, from data acquisition to post-processing, that not only result in a noisy appearance to the CHM but may also limit semi-automated tree...
We introduce an automated change detection and delineation tool for remote sensing images: the Land-cover Change Mapper (LCM). LCM rapidly generates a polygon vector layer (shapefile) of regions deemed to have undergone significant change in land-cover. In its simplest usage, LCM requires two single band or multi-band co-registered images of the sa...
Forest monitoring information needs span a range of spatial, spectral and temporal scales. Forest management and monitoring
are typically enabled through the collection and interpretation of air photos, upon which spatial units are manually delineated
representing areas that are homogeneous in attribution and sufficiently distinct from neighboring...
KEY WORDS: Canopy height, Geographic object-based image analysis (GEOBIA), Geographic object-based texture (GEOTEX), Tree-ray-shadow geometry (TG), Quickbird, Lidar ABSTRACT: Canopy surface height (CSH) is a significant forest biophysical parameter to estimate above-ground biomass and carbon content. High-spatial resolution optical remotely sensed...
The manual delineation of vegetation patches or forest stands is a costly and crucial stage in any land-cover mapping project or forest inventory based upon photointerpretation. Recent computer techniques have eased the task of the interpreter; however, a good deal of craftsmanship
is still required in the delineation. In an effort to contribute to...
High spatial resolution satellite imagery, with pixel sizes of one metre or less, are increasingly available. These data provide an accessible and flexible source of information for forest inventory purposes. In addition, the digital nature of these data provides an opportunity for automated and computer-assisted approaches for forest stand delinea...
Object-Based Image Analysis (OBIA) has gained considerable impetus over the last decade. However, despite the many newly developed
methods and the numerous successful case studies, little effort has been directed towards building the conceptual foundations
underlying it. In particular, there are at least two questions that need a clear answer befor...
What is Geographic Object-Based Image Analysis (GEOBIA)? To answer this we provide a formal definition of GEOBIA, present
a brief account of its coining, and propose a key objective for this new discipline. We then, conduct a SWOT1 analysis of its potential, and discuss its main tenets and plausible future. Much still remains to be accomplished.
What is Geographic Object-Based Image Analysis (GEOBIA)? To answer this we provide a formal definition of GEOBIA, present a brief account of its coining, and propose a key objective for this new discipline. We then, conduct a SWOT¹analysis of its potential, and discuss its main tenets and plausible future. Much still remains to be accomplished.
Forest monitoring information needs span a range of spatial, spectral and temporal scales. Forest management and monitoring are typically enabled through the collection and interpretation of air photos, upon which spatial units are manually delineated representing areas that are homogeneous in attribution and sufficiently distinct from neighboring...
Object-Based Image Analysis (OBIA) has gained considerable impetus over the last decade. However, despite the many newly developed methods and the numerous successful case studies, little effort has been directed towards building the conceptual foundations underlying it. In particular, there are at least two questions that need a clear answer befor...
This paper deals with the description and assessment of uncertainties in land use data derived from Remote Sensing observations, in the context of hydrological studies. Land use is a categorical regionalised variable reporting the main socio-economic role each location has, where the role is inferred from the pattern of occupation of land. The prop...
Large area thematic mapping is usually performed through the digital classification of satellite imagery. Regardless of the method selected, pixel-wise classified images typically suffer from 'salt and pepper' noise that hinders raster to vector conversion. In order to enable a correct vectorization, classified images require a post- processing ste...
The manual delineation of vegetation patches or forest stands is a costly and critical stage in any landcover mapping project or forest inventory based upon photointerpretation. Recent computer techniques have eased the task of the interpreter; however a good deal of craftsmanship is still required in the delineation. In an effort to contribute to...
Los principios que inspiran la futura Directiva Europea sobre Infraestructuras de Datos Espaciales han dado lugar a la puesta en marcha de un nuevo Sistema de Información sobre Ocupación del Suelo de España (SIOSE), donde se pueda integrar la información de las Bases de Datos de Ocupación del Suelo de las Comunidades Autónomas y de la Administració...
As an emerging discipline, we propose a formal definition of OBIA, describe how OBIA came into existence, and as a road map to future research propose a fundamental objective. In order to provide potential strategies to meet this objective, we undertake a tentative SWOT Analysis to identify current Strengths, Weakness, Opportunities and Threats tha...
Over the last decade the analysis of Earth observation data has evolved from what were predominantly per-pixel multispectral-based approaches, to the development and application of multiscale object-based methods. To empower users with these emerging object-based approaches, methods need to be intuitive, easy to use, require little user interventio...
Thematic maps represent the spatial distribution of phenomena such as vegetation, land use or geology. Given a level of both cartographic and thematic detail , errors can be defined as deviations between the representation and the represented terrain that can not be explained by the spatial and categorical generalization mechanisms that were applie...
This paper describes progress on HarmoniRiB, a European Commission Framework 5 project. The HarmoniRiB project aims to support the implementation of the EU Water Framework Directive (WFD) by developing concepts and tools for handling uncertainty in data and modelling, and by designing, building and populating a database containing data and associat...
Se presenta un método para obtener de forma automática, a partir de una ortoimagen, una malla de recintos irregulares de tamaño superior a un cierto mínimo. Esta malla es similar a la que delinearía un fotointérprete al que se le diera la tarea de dividir la imagen en regiones relativamente uniformes en su interior y diferentes de lo que las rodea,...
Landcover maps typically represent the territory as a mosaic of contiguous units –polygons-that are assumed to correspond to geographic entities –like e.g. lakes, forests or villages-. They may also be viewed as representing a particular level of a landscape hierarchy where each polygon is a holon –an object made of subobjects and part of a superob...
Object-oriented analysis of RS images for landcover mapping is based upon the same hierarchical patch model used in modern Landscape Ecology. In such model, each patch is a loosely integrated whole -- an object that can be viewed simultaneously as part of a superobject and as made of subobjects. The focal level, i.e. the level of the nested hierarc...
Complete manual-guide of natural parameters using for environmental studies and ecological planning
Object-oriented analysis of RS images for landcover mapping is based upon the same hierarchical patch model used in modern Landscape Ecology. In such model, each patch is a loosely integrated whole -- an object that can be viewed simultaneously as part of a superobject and as made of subobjects. The focal level, i.e. the level of the nested hierarc...
In the last decades, there has been a trend towards automated landcover mapping.
Nevertheless, the overall accuracy of the maps produced in this way is normally below the
user’s requirements. Consequently most landcover mapping projects still rely to some extent
on photointerpretation, which is less cost-effective and more subjective than the forme...
Questions
Questions (5)
Today I came up with a document that used the term satellite-based earth observation (SBEO). I find the term is highly redundant, because in my mind, Earth Observation (EO) is the subset of remote sensing that is done from space. Then I read this in the Wikipedia article on EO:
"The term Earth observation is used in two ways, leading to confusion. In Europe, in particular, it has often been used to refer to satellite-based remote sensing, but the term is also used to refer to any form of observations of the Earth system, including in situ and airborne observations, for example. The Group on Earth Observations, which has over 100 member countries and over 100 participating organizations, uses EO in this broader sense.
To add to the confusion, in the US, for example, the term remote sensing is often used to refer to satellite-based remote sensing, but sometimes used more broadly for observations using any form of remote sensing technology, including airborne sensors and even ground-based sensors such as cameras. Perhaps the least ambiguous term to use for satellite-based sensors is satellite remote sensing, or SRS, an acronym which is gradually starting to appear in the literature."
Have you ever used the term SRS, and do you think SRS should be favored over EO, or what else can be done to reduce confusion?
NB. I use 'spaceborne' instead of 'satellite' for the 1st 'S' in SRS because the International Space Station (ISS) is used as remote sensing platform and most people would not consider the ISS a 'satellite'.
I have used in the past both "remote sensing" and "remotely sensed" as qualifiers to refer to data acquired by airborne or spaceborne sensors. However, I feel increasingly uneasy about "remotely sensed", because I find it convoluted and even a bit pedantic. I would like to convince co-authors to give up using "remotely sensed", and I offer two arguments:
- A digital picture of a lake taken from the top of a mountain was indeed remotely sensed, but it's clearly not what we would understand as a remote sensing image.
- 'remotely sensed' is a secondary jargon derived from the primary jargon 'remote sensing'. Why should we use yet one more jargon when the original one can also work as an adjective?
Can you come up with any other supporting argument, or if you prefer 'remotely sensed', can you come up with a counterargument?
If I tell you that I used 48 Landsat scenes in a study, what would you think I'm referring to: 48 different WRS2 Path-Row combinations, 48 images of the same Path-Row covering several years, or maybe 6 different Path-Rows with 8 acquisition dates each?
Although officially 'scene' refers to the individual data frame into which the satellite continuous data streaming is segmented, and thus it applies to the image data, I think we would be better served if 'scene' was reserved to refer to the area of the earth imaged in that path and row. That is, it would be good if there was a consensus that 'scene' is a synonym for 'tile' or 'map sheet' in the context of Landsat, so that folks know unambiguously what I'm referring to when I say that I used 48 Landsat scenes in a study.
What do you think?
We found in an object-based accuracy assessment of a landcover polygon map (doi:10.1080/01431161.2013.875630) that the likelihood of correct classification of patches decreases as their size decreases, slightly for large and medium-sized patches, and more dramatically for patches within two orders of magnitude of the pixel size of the input image. Whenever the correct class of a misclassified patch is shared by a correctly classified adjacent patch, the local configuration of patches will be misrepresented in addition to the local composition of classes. Therefore, if further studies confirm the abrupt decline in accuracy we observed, the suggestion by Langford et al. (2006; Ecosystems 9: 474–488) that “there is potential for large errors in nearly every landscape pattern analysis ever published” would become an understatement. Are you aware of any studies beyond Smith et al. (2002; PE&RS 68:65-70) supporting (or contradicting) this finding, and do you think that if confirmed, this would be reason enough to abandon the patch-mosaic paradigm?
NB. For those lacking institutional access, a free copy of my article “The impact of object size on the thematic accuracy of landcover maps” can be downloaded from:
until the download allocation (50) is exhausted. Please do not use if you can download through your institution.
I'm writing a paper about this subject and would need to know if there are tools out there that can perform the following task: extract a random subset of polygons from a vector layer whose total area would be equal to a user-defined sampling intensity. Please note I don't need a recipe on how to do this (I've actually created such a tool), I just need to know what others tools are already implemented in commercial or free GIS packages that can do this directly.
