Thoughts on Technology

Timothy Killeen, University Corporation for Atmospheric Research

I became the Director of the National Center for Atmospheric Research or NCAR ( in July 2000, after a 23-year career as a professor and administrator at the University of Michigan. NCAR is the only national center dedicated to the integrated study of weather, climate, atmospheric chemistry, solar and space physics, and the socio-economic impacts of natural and anthropogenic environmental change. NCAR and its governing body, the University Corporation for Atmospheric Research, UCAR, have a strong education and outreach program ( that ranges from informal science education (e.g.,, to teacher training workshops, undergraduate mentoring programs for underrepresented minority students ( and graduate and postdoctoral scholar programs.

At the University of Michigan, I was a teaching faculty member of the Atmospheric Sciences department in the College of Engineering and Associate Vice-President for Research with some responsibility for the integration of research and education in the undergraduate programs across campus. In this latter position, I organized the 1999 Jerome K. Wiesner Symposium entitled "New Integrations of Research, Scholarship, and Undergraduate Education" ( I was also involved in administration of the award-winning Undergraduate Research Opportunities Program (UROP) and the Women in Sciences and Engineering Residential Program (WISE-RP) at Michigan. I also helped to organize and lead the first annual Michigan Institute on Integrating Undergraduate Research and Education ( in the summer of 2000.

Under NSF sponsorship from the Institution-Wide Reform of Undergraduate Education (IR) program, I led the interdisciplinary faculty team that, over 8 years, developed the multi-semester course sequence in Global Change ( This course sequence forms the core curriculum of the first "front-loaded" interdisciplinary minor to be offered at Michigan. The term "front-loaded" here means that the minor is open to all students and can be completed in the first two years at the university, before a major needs to be selected. The sequence has been the subject of an extensive formative and summative evaluation effort that has documented the efficacy of the IT innovations used in both the laboratory and classroom portions of the curriculum.

Recently, I was asked to speak at the 2001 "Workshop on the Impact of Information Technology on the Future of the Research University: Launching the Dialog", sponsored by the National Academies' Government-University-Industry Roundtable The following paragraphs are taken from my short oral contribution and reflect my current thinking about IT in the undergraduate setting.

Let me give you a brief summary of professional experiences I have had which inform my views on the impact of IT on the teaching, research, and service activities of Research Universities.

First, what is NCAR? NCAR is an > $100M/year FFRDC (in fact funded by NSF's largest single cooperative agreement) with a mission of research, service, and education in the atmospheric and earth system sciences. NCAR is a supercomputer center, with a large mass store and fiber optic connectivity - it has a fleet of research aircraft, large and small radars, 9 scientific divisions and a stellar history of contributions to both the geosciences and to computational science research.

Before I joined NCAR in June of last year, I was Associate Vice-President for Research at the University of Michigan and a long-term faculty member in the Engineering School. I have carried a full graduate and undergraduate teaching load for the past 15 years, as well as running a research group involved in building satellite instruments and analyzing earth remote sensing data. Part of my role in the central administration was to foster the effective integration of research and education through the use of IT. I led a project for the "Institution-wide Reform of Undergraduate Education" at Michigan that was based on the use of IT to develop pilot interdisciplinary, introductory courses for non-science majors - more on this later. I was also involved as the science lead for a long-running Collaboratory project dealing with the physics of the near-earth space environment.

So, you might say that I have been steeped in this arena for some time and from all the perspectives of full time faculty, researcher, university administrator, and now national lab director.

My three main points are:
1. That IT in the research university can - and indeed should - seed a national re-interpretation and re-design of general education. I will focus here on the scholarship of learning and specifically the issue of scientific education for non-science majors

2. That much of the needed knowledge generation for the next few decades requires the efficient deployment of the interdisciplinary expertise of the research university - importantly including the arts and humanities.

3. That investments are urgently needed now to develop and implement the computational science underpinnings to enable the kind of public interoperability and access that will be needed to extract and share added value content from our data-rich and high-bandwidth future. Collaboratories, common modeling frameworks, and meta-data standards for digital libraries are all examples of areas where investments will pay off.

Let me start with some comments to illustrate my first point - the need for a re-invention of general education.

The need for effective science, mathematics, engineering and technology (SME&T) teaching and learning at the introductory undergraduate level is one of great import and scale. Consider that, in the U.S. postsecondary educational system alone, there are roughly 14 million students enrolled in 3,600 institutions. These students earn 1.9 million degrees per year of which about 1.4 million are granted in non-science and engineering areas. These students (both science and non-science majors) are the teachers, legislators, industrial decision-makers and researchers of tomorrow. They will need a working background and knowledge of science to confront the complex challenges of an increasingly technological society in a world of limited natural resources. How well are we doing? - not well enough!

It is clear that we cannot afford a postsecondary educational system that "turns off" students from even a rudimentary appreciation for scientific thought and quantitative analysis. Yet, there are indications that this is exactly what is happening throughout the educational system. The findings of the Third International Mathematics and Science Study (TIMSS) showed that only one quarter of U.S. high school students enroll in physics and only one half in chemistry. It pointed towards curriculum materials and pedagogical approaches that were unsuccessful in capturing student interest.

At the undergraduate level, there are similar indicators of a systemic failure to capture the interest of students who are not (at least initially) inherently motivated to study SME&T topics. A study in 1995 indicated that fewer than 20% of students take a SME&T course after their freshman and sophomore years. College attrition rates are very high. Of the 2.4 million students entering 4-year colleges in 1993, 1.1 million left without a degree. The figures are lower for specific underrepresented groups. The central concern here, of course, is that fewer and fewer citizens are comfortable with the concepts of science and technology at a time of greatly expanded societal reliance on such tools. This is a concern that must be shared among all faculty and administrators in higher education.

It is one that IT can help resolve. And here, the Universities must help lead.

There is a growing body of evidence from what some call the "Scholarship of Learning" that the effective integration of research and education is one answer to the student engagement problem. Undergraduate Research Opportunity Programs, for example, are showing significant progress with measurable gains in GPA levels and retention rates for students entering them. But, there is a problem of scale - at the University of Michigan, for example, only about 900 of 35000 students are in the program, due to natural limitations of faculty time and other factors. The problem for research universities therefore is to distill the essence of these richly mentored research experiences - what makes them work - and use this information to understand how to "wholesale" research-enriched learning in the regular curriculum, without dilution in quality. This is where the scholarship of learning - and IT - come into play. Through effective instructional technologies based on IT, engaging, accessible, interdisciplinary introductory courses that are science-based can be assembled, modularized, maintained, evaluated, shared, and subjected to continuous improvement. IT can be used in the many ways that you are all familiar with: to cross disciplinary boundaries seamlessly, to conduct virtual field trips, to support reflective asynchronous study with feedback, to carry forward exciting synchronous (but distributed) team work, to access and manipulate remotely-held data sets, to model complex systems, to address and appreciate the meaning of uncertainty - and so on. Students can and do become engaged with science in this way.

We demonstrated this approach at the University of Michigan in a small way, where we conducted a successful 8-year experiment with a truly interdisciplinary course sequence in global change, integrating materials from natural and social sciences and the humanities. Our interdisciplinary team involved professors from 8 departments and 5 schools who met weekly for over five years with a team of graduate students, postdocs and evaluation experts to build an IT-enriched multi-semester curriculum. Through the extensive use of IT - web-based materials, dynamical modeling tools, interactive data analysis tools, evaluation instruments, tailored distance learning modules, etc., the course sequence became the most "interesting to students" of any introductory science-based course on campus and independent evaluations (conducted by the School of Education) showed that alumni of the course had a greater propensity for taking advanced courses and going on to graduate school.

To give you just a flavor of what our non-science majors were doing - After 6 weeks of class, students with very little or no advanced math background were using dynamical modeling tools to study global warming scenarios of their own devising for Earth and Venus - and assessing the likelihood of a future ice age by analyzing and manipulating data from disparate sources, including ice core and other paleoclimate records, isotopes in the ocean, atmospheric carbon dioxide and methane trends, Milankovitch cycles describing changes in the earth's orbit around the sun, etc. All data are pulled over the web, inserted into dynamical modeling and graphical analysis packages and studied in the context of a small-enrollment lab sections. A few weeks later they were conducting a modeling study of how fish populations in the Outer Banks respond to regulatory action of different types - and later still quantitatively studying the impact of world-bank loans on human population mobility due to dam building in the far east.

Now the three-course global change sequence is the basis for a brand new interdisciplinary minor degree at the university - one that students complete in their first two years. The minor provides a new model for general education - it puts interdisciplinarity before disciplinarity (some referees have great problems with that!) - it gives students access to numerous professors from multiple departments early on before the student declares a major. It also gives students a more profound appreciation for the bi-directional relationship between humankind and the planet. In that sense, it's not "pre-Med", but more like "Pre-Life."

But I have to admit that it is not a trivial matter to experiment profoundly with the use of IT in university curricula and, in fact, just getting this on the books originally took two years! However, such new models of general education are springing up in most research universities and are being evaluated for effectiveness in an exciting national reform movement, all spurred forward by IT innovations.

The point I want to make here is that we, as a national community of educators, have learned something about how IT can be used to improve the educational experience. With future bandwidth and computational improvements, and if the research universities help take the lead - further gains can be made, ultimately having a profound impact on the scientific literacy of the our citizens and fulfilling the promise to parents of an education for their children truly enriched by the research environment.

My second point really speaks to the essential contributions of the research university for scientific progress in high priority areas. And once again, I'd like to speak very briefly to the area of my own research interests, but I think the comments have greater generality.

The overarching purpose of earth system science is to develop the knowledge basis for predicting and adapting to future changes in the coupled physical, chemical, geological, biological, and social state of the earth and assessing the risks associated with such change. I would argue that the most natural engine for such an interdisciplinary research agenda is the research university and that the only way to take this on is using extensive IT tools. No think tank has the disciplinary range, no company can have the long-term motivation or resources to address such a set of nested challenges alone. Yet, as we all know, there is a clear societal imperative for such study. I could list many impending changes, but let's just focus on one - during this century, humankind will essentially become an urban species, with over 4 out of every 5 people living in such settings. The implications of this for atmospheric chemistry, for energy policy, for transportation, for surface hydrology, etc. etc. are enormous and require deep interdisciplinary study - including the efforts of the arts and humanities whose role it is to describe and celebrate the human condition.

So, how can such studies be carried out? Through observation, theory and modeling (the so-called "third modality"). Since these are very complex systems, integrated teams of mathematicians, economists, geoscientists, computer scientists, artists and humanists will be needed and significant knowledge-generation, dissemination, and curation capabilities will be needed.

In the next 5-10 years, the data rate of information describing the earth's state coming from satellites will increase by more than three orders of magnitude. We are not ready for this. In my own institution, we are dealing with terabytes of new data each week and will soon be confronting a petabyte data curation problem. The coupled models of the earth system are now truly community artifacts - not owned (or even understood) by single individuals, but rather by the broad community of practice. For example, NCAR's community climate system model (CCSM) is being co-developed by over 20 universities and its governance structure involves 13 separate working groups.

My point here is simply this - the scientific and intellectual challenges posed by the earth system sciences and the human dimensions of global change simply require a full-court press of IT capabilities as well as the harnessing of the interdisciplinary expertise base developed and sustained in the research university.

My last point is quickly made. The challenges that I have spoken of - the need to reform education for non-science majors and the need for interdisciplinarity in the earth system are just two examples of a broader set of issues. They all have in common the need to generate, disseminate and curate knowledge digitally. For universities to be able to fully realize their potential, we need some rules - albeit flexible ones - for this game. Just as libraries became much more powerful when cataloging systems were standardized, the process of knowledge generation has to have some kind of mutually understandable set of semantics. In my work with collaboratories, I have seen how the buy-in is greater, and therefore the net benefit is greater, when the learning curve for involvement is not perceived as a major barrier. Knowledge systems involving digital libraries, collaboratories, data and data transport systems and large scope models should all be designed from the outset to be generalizable, scalable, and sharable, so that not too many wheels are re-invented. These goals are non-trivial and will require careful attention and resources. Now is the time to make these investments carefully in IT infrastructure.

In summary, I am a true believer in the transformational character of the IT revolution for the university - they are simply made for each other. IT will change general education in profound ways - and we are only just beginning to appreciate by just how much. It is certainly true that university administrations will need to show flexibility and imagination. IT will also enable us to learn about nature - and to share the excitement of learning about nature - in ways that are intrinsically interesting, rigorous and relevant to life on earth. What more could we ask from our computer science friends?