Professor Stefan Leyk received a $499,999 grant for the Imola project from the National Science Foundation (NSF) for the titled project, . Collaborators on the project include Associate Professor from the University of Minnesota. 

The Imola (Intelligent Map recOgnition LAb) project develops advanced computational methods and scientific approaches to extract historical geographic information from scanned maps originally published by the US Geological Survey between 1884 and 2006.  The project transforms historical road networks in these maps into a database that allows scientists to study the evolution of transportation networks and how humans interact with their environment over extended time periods and geographical regions prior to the era of satellite imagery.

Extracting detailed geographic information from historical USGS topographic maps is a difficult task. The Imola project uses a new intelligent system called DaVinci to automate the extraction process and produce a large spatiotemporal database of historical road networks called US1884+.  DaVinci reads maps like humans by automatically exploring geographic features in maps and in other datasets to generate more information related to historical road networks to then extract them. The DaVinci method eliminates the need for large, manually created training data and provides a way to measure uncertainty in the features extracted from the maps, which is important for conducting research with the data. The project provides a case study and creates tutorials to demonstrate how the US1884+ platform can be used in scientific research and how it advances knowledge of how humans interact with their environment.

Off

Johannes Uhl

Nature published a  today, by Johannes Uhl, about the team's work.

• Uhl J.H., Leyk S., McShane C.M., Braswell A.E., Connor D.S., and Balk D. 2021. Fine-grained, spatio-temporal datasets measuring 200 years of land development in the United States, Earth System Science Data, 
• Uhl J.H., Connor D.S., Leyk S., and Braswell A.E. 2021. A century of decoupling size and structure of urban spaces in the United States, Nature Communications Earth & Environment, 

[video:https://youtu.be/Le7mVCyeeVQ]

[video:https://www.youtube.com/watch?v=__eGCxLzX6o]

[video:https://www.youtube.com/watch?v=4efdASKNNMA]

[video:https://www.youtube.com/watch?v=jyCz_ON8Z4M]

Off

The U.S. population has grown 40-fold since 1800 and yet little is known about precisely how or why it’s grown the way it has. New research from the University of ŷ���ڱ���Ƶ Boulder, though, provides the most comprehensive look yet.

This new research, , draws a detailed picture of how and why the United States formed the way it did, improving our understanding of the history and evolution of human land-use and development in this country, and offering new opportunities to understand the spatial distributions of past populations.

Building on previous work by lead author Stefan Leyk and Johannes Uhl, these new visuals and data also provide an unprecedented chance to study long-term processes that could have implications on our understanding of future risks like sea level rise and ŷ���ڱ���Ƶ, as well as the rural-urban divide.

“I’m a very strong believer that we need to know more about the past, what we did in the past, how our landscape has evolved in the past, and what, specifically, it means when we look at the built environment, to better understand the past 100 or 200 years of urban development,” said Leyk, an associate professor of geography at ŷ���ڱ���Ƶ Boulder. 

“We’re trying to understand the history, understand the evolution, understand how this socio-environmental system actually works …  Because there are so many different components to it.”

Between 1800 and 2015, the population of the contiguous United States grew from 5.3 million to 309 million. During that time, humans sprawled out across the continent but researchers’ understanding of that growth has been disjointed at best.

Historically, there have only been two ways to study the change and evolution of landscapes: population- and land-based methods. Population-based methods rely on census records, but the lack of consistency over time and detailed records before the mid-20th century means that this method is challenging and intensive for researchers. Land-based methods (which look at historical land-use or how the land was used), on the other hand, are often limited to a few decades, and are incomplete or inaccurate for specific characteristics of built-up land or for rural areas.

Together, these approaches create an uneven picture of how the United States developed, increasingly becoming muddier when you zoom in and use finer scales of time and space.

Which, according to Leyk, is a persistent problem that geospatial scientists across the globe are trying to fix.

For Leyk and colleagues, the solution to that problem for the United States came from an unlikely ally: Zillow.

They worked with the housing company to get permission to use its massive dataset, the Zillow Transaction and Assessment Dataset (ZTRAX), which contains more than 374 million data records covering every single property it has on record. The researchers, with the help of funding from the National Science Foundation and National Institute of Health, then used that to bring both land- and population-methods together to create a new, unique source: The Historical Settlement Data Compilation for the United States (HISDAC-US). 

“It’s a big deal to bring these two perspectives together,” Leyk commented.

With this new settlement layer dataset, which is free for anyone to use and , researchers mapped the built environment of the conterminous United States all the way back to 1810 at both fine temporal (five years) and spatial (250 meters) granularity.

Using these new data, researchers were not only able to see well-established land-use trends—such as urban development going westward over time or some areas building up before they joined the United States—they were also able to determine detailed growth patterns for human settlements, such as what factors may cause cities to expand or densify or how the growth of urban settlements compares to rural settlements.

For Leyk and his colleagues, though, this research is only a beginning. They are already using these historical data on several other projects—one such being better understanding the future threats that natural hazards may pose by looking at how we got here.

“It’s a big deal how we, as people in our built environments, are and have been exposed (to natural hazards) over long periods of time. We don’t know much about that, but the historical trajectories will tell us a lot about these changes, such as along the coastline.”

Other authors of the paper include Dylan S. Connor from Arizona State University; and Uhl, a former geography graduate student, Anna E. Braswell and Nathan Mietkiewicz, research scientists with Earth Lab, Jennifer K. Balch, director of Earth Lab and a geography professor, and Myron Gutmann, a history professor and the director of the Institute of Behavioral Sciences, from ŷ���ڱ���Ƶ Boulder.

Off

• Message from the chair
• John Pitlick: Guggenheim in the rearview mirror
• Stefan Leyk: Extracting Geospatial Data from Historical Maps 
• Fernando Riosmena: Is Climate Change A Driver of Migration?
• Dylan Lanka: Solifluction Erosion
• Kai Kresek: Perception of Neighborhood Space
• Earth Lab Launches New Professional Certificate in Earth Data Analytics
• Department News

All previous newsletters are on our Newsletters page.

For a more enjoyable reading experience, open a newsletter file and adjust your browser window to the same size as the newsletter page. The Table of Contents and other links are active within the document. Clicking on the black box at the top of each page returns the view to the Table of Contents.

Off

Ph.D. student Johannes Uhl attended the symposium in person and gave a talk today providing technical details on the research.

Off

Do you want to enter the job market as a competitive GIS modeler with programming skills? This course will help you get there. It focuses on the extension of geographic information systems (GIS) through programming as well as on the development of algorithms for spatial analysis and information extraction in vector and raster data. We will cover different concepts, principles and techniques of programming that help you to solve a variety of spatial problems in physical and human Geography. You will learn how to work with Python for Geoprocessing in ArcGIS as well as for spatial programming in gridded data using numpy, scipy and other open source libraries. Furthermore, you will understand the basic ideas of object-oriented and procedural programming. You will develop skills to explore, handle, manipulate, and model spatial data as well as methods development. Lectures will include numerous demonstrations and hands-on examples as well as algorithmic exercises. In labs you will work on solving typical programming and implementation problems that you will encounter in the real world. During the last weeks of the term students will work in small groups on a proposed project to deepen their programming knowledge, improve their GIS proficiency and train their presentation and communication skills.

Off

Did your first GIS class create an interest in learning more advanced skills? Are you getting ready to start an undergraduate or graduate research project that requires spatial modeling and analysis, but you’re not sure how to get started? This class is for you!  You’ll extend skills and principles, gain confidence in your GIS knowledge, and get hands-on technical experience with the full spectrum of GIS modeling. You will learn to implement line-of-sight models, proximity models, design hydrologic and terrain analyses, and work with point interpolation and kriging, dasymetric small area estimation, weighted criteria estimation, sensitivity analysis, and modeling landscape change. You will learn best practices for exploring data and using models to search for patterns that will help you in many areas of physical and social geographic analysis and environmental modeling. You'll work with raster data in a visual programming environment, learning automation and GIS scripting methods that will help you in your own research or in a job or internship. This class provides an excellent transition between Introductory GIScience (GEOG 4103/5103) and GIS Programming (GEOG 4303/5303).

The class format includes lectures, weekly in-class demos and exercises, and weekly lab assignments, each on a specific modeling task. In the second part of the semester, students will work in small groups to design, run and evaluate GIS models for projects they choose, with guidance from the instructor and TA. Each group will present project results to the class and complete a report to be handed in the final week of term. Project leaders will design and manage the project and the group activities, learning basic project management skills by working one-on-one with the instructor.

Off

Associate Professor Stefan Leyk has been awarded a National Science Foundation grant for his project entitled "III: Medium: Collaborative Research: Exploiting Context in Cartographic Evolutionary Documents to Extract and Build Linked Spatial-temporal Datasets". This interdisciplinary collaborative project involving researchers and their students at University of Southern California and University of ŷ���ڱ���Ƶ, Boulder will develop a set of open-source technologies and tools that allow users to extract map features from a large number of map sheets and track changes of features between map editions in a Geographical Information System. The resulting open-source tools will enable exciting new forms of research and learning in history, demography, economics, sociology, ecology, and other disciplines. The data produced by this project will be made publically available and through case studies integrated with other historical archives. Spatially and temporally linked knowledge covering man-made and natural features over more than 125 years holds enormous potential for the physical and social sciences. The wealth of information contained in these maps is unique, especially for the time before the widespread use of aerial photography. The ability to automatically transform the scanned paper maps stored in large archives into spatio-temporally linked knowledge will create an important resource for social and natural scientists studying global change and other socio-geographic processes that play out over large areas and long periods of time.

Off

Stefan Leyk has received a new award from the Innovative Seed Program, “Earth Lab’s human dimension: Integrating fine-grained data on human activity for advanced understanding of environmental change.” It will become a new node in Earth Lab, one of two major initiatives of ŷ���ڱ���Ƶ’s Grand Challenge.

See for more information.

Off

Stefan Leyk and Fernando Riosmena are part of a research team that received a $1 million grant from the National Science Foundation to conduct research on the distribution and dynamics of the world’s population, including the modelling of urbanization in the United States, Mexico, and India.

Off