Ocean Modeling Discussion

The oceanography community has lost a respected pioneer of regional and coastal ocean modeling. Dale Haidvogel passed away unexpectedly on March 14 at his home in Buffalo, NY. Though recently retired from Rutgers University, Dale was still active as an emeritus instructor in graduate GFD courses.

A graduate of the MIT/WHOI Joint Program, Dale subsequently held scientist and faculty positions at WHOI, NCAR and The Johns Hopkins University Chesapeake Bay Institute before being recruited to Rutgers in 1990 to lead an ocean modeling group in the newly formed Institute of Marine and Coastal Sciences.

Dale was an advocate for introducing advanced numerical methods into ocean simulation and was instrumental in coordinating early efforts at rigorous inter-comparisons of differing model codes. He was especially passionate about spectral methods and the joint analysis of models and laboratory tank experiments of fundamental processes in geophysical fluid dynamics. He inspired efforts on coupled circulation-biogeochemistry simulation and was a long-time member and chair of the Scientific Steering Committee of U.S. GLOBEC (Global Ocean Ecosystem Dynamics) program.

Like many others in the Rutgers Ocean Modeling Group extended family, Dale was instrumental in setting me on my professional trajectory. In 1985 I began working with "Dale's model" as part of my PhD thesis research. The model evolved into SPEM (the Semi-spectral Primitive Equation Model) and echoes of the SPEM source code permeate ROMS to this day.

Dale’s kind and cheerful approach to life and his profession will be missed. We all feel this loss acutely.

I met Dale when I was a PostDoc at Harvard University. He came to our group and gave an excellent presentation about ocean modeling. We also met in my office to discuss SPEM, which I had been learning. When I was ready to leave Harvard, I emailed him about my chances for me to get a modeling position at NCAR as my next move in my professional career. He called me the next day and invited me to come and visit Rutgers. He, Scott Glenn, and Fred Grassle convince to go to come to Rutgers instead in the fall of 1995. My first task was to get better solutions from SCRUM. In particular, he and Kate Hedstrom were interested in the Big Bad Basin test case (CPP BASIN). The classic double gyre solution that was failing in SCRUM. I rewrote that model from scratch and became SCRUM 4.0, which evolved to ROMS when I started working with our collaborators at UCLA, Sasha Shchepetkin and Jim McWilliams. The rest is ROMS history.

He was a family man passionate about cooking and very hot food.

I remember fondly discussing the new ROMS model test cases in his book and the three papers for Dynamics of Atmospheres and Ocean (Haidvogel et al., 2000; Malanotte-Rizzoli et al., 2000; Chassignet et al., 2000). Those are the first publications mentioning ROMS.

Dale will be sorely missed.

Dale Haidvogel played a big part in getting my career started. My PhD project involved a rotating tank with actual salt water. My first job afterwards was to do ocean modeling, but I didn't know anything about it - there were no classes in it then at Scripps. My new boss at the (now long gone) Institute for Naval Oceanography (INO) asked several of his friends if they would teach me. Dale Haidvogel was the one who agreed, in exchange for having me write the first SPEM manual. I spent three months in his lab in Baltimore, learning what I could before heading to Mississippi to learn LaTeX and write it all up. We got along well and I continued to work with Dale and with SPEM until things changed at INO. It was then that I asked Dale about some support positions at NCAR and he suggested that I come join him instead. He knew he'd be leaving Balitmore, but he didn't know where his next position would be. I agreed to join him in Baltimore, then go wherever. I was with him at Rutgers for ten years, during which we had a lot of fun challenges with new computers, with sea ice modeling, ocean model test problems and with grid generation. His teachings live on in my suite of test problems for MOM6 open boundaries.

Dale was my postdoc advisor, and he was the one who introduced me to ocean modeling which I've been doing ever since. I came to Rutgers in 1995, and started working on Spectral Element Ocean Model (SEOM) with him and Mohamed Iskandarani. Dale was very involved in this project, and I have fond memories of us spending hours discussing various aspects of model formulation. Over the years we worked on many projects together, both with SEOM and ROMS, doing global and regional modeling of physical, and later, coupled physical-biological processes. I learned a lot from him. He was always approachable, kind, and unassuming, and I am very saddened by this loss.

Julia Levin

I'll echo everyone's heartfelt comments about Dale. Though as a graduate student I knew his papers, I first got to know Dale in the late 1990s when I was a postdoc on a US Globec project with Zack Powell and him. I traveled to Rutgers often (working then with SEOM) and enjoyed many detailed modeling and other discussions with the group that was active there at that time (Dale, Hernan, Kate, Mohamed, Julia, Enrique, John). Dale's enthusiasm for ocean physics and multidisiplinary problems and ocean modeling and numerical methods was infectious and inspiring for a junior scientist. I enjoyed his book and professional presentations as well (I remember a wonderful talk at a Gordon Conference on Coastal Ocean Modeling on scaling laws appropriate for tidal dynamics over canyons driving cross-shelf exchange that linked nicely to some rotating table experiments). I learned a lot from Dale and appreciated his boundless humor, optimism, and kindness.

Chris Edwards

It was a shock with great sadness to learn about the loss of Dale Haidvogel. He was a great mentor and an inspiration to my life and my career.
I still vividly remember my time working with him at Rutgers 30 years ago. I first met him at a NCAR workshop, where he accepted me as his postdoc. He was a few pioneers in ocean modeling then. At beginning, I was learning his SPEM, with great help from Kate Hedstrom.
He was so kind and always encouraging me to do independent research, never demanded anything. A “free-hand” allowed me to create the “s-coordinate” for SCRUM, the idea I borrowed from Econometrics. Though called stretching-level, the “s-curve” was actually describing a company from taking off to aging out. To my surprise, it works for resolving ocean’s surface and bottom layer, generalized the sigma-coordinate without extra computing cost.

Before his retirement, he visited JPL yearly.
He was quite healthy then, and often played Ping-Peng with me and colleagues at my home. Life is so short.
He will surely be missed!

Y. Tony Song

I remember meeting him at a conference at Rutgers, after which he invited everyone to his house. He was one of the first "big names" that I met, and realized that they were not just names on papers but actual people -- and idea that was still new to me, and a bit of a shock. He was very kind in explaining to me how SPEM worked when I was a student. I have always tried to model how I have answered random questions out of the blue after how he took time to answer my questions -- clueless and out of the blue as my questions were.

Jamie

I too have many fond memories of my interactions with Dale over the years. I particularly remember him as kind, supportive and enthusiastic.
It was Dale who brought Hernan Arango, Art Miller, Bruce Cornuelle, Manu Di Lorenzo and myself together to discuss how we might develop the ROMS tangent linear and adjoint codes.

Dale will indeed be greatly missed by us all.

Andy

I am saddened by the news. Dale Haidvogel was the person who brought me to the U.S. and into the oceanography community, and let me pursue my own interest. For that, I am forever grateful!

I was in Fluid Mechanics when I was working on a master's degree at Zhejiang Univ in China 20 years ago. By accident, I discovered my interest in oceanography. So, I emailed a number of professors in oceanography in US asking for Ph.D. student opportunities when I applied for graduate schools. Dale was the only one who responded. After I submitted my application to Rutgers, Dale set up a phone interview with me. I remember being extremely nervous while waiting in the dorm room for his call. I could barely speak English at that time. Fortunately, Dale did most of the talking on that call, and he didn't really ask me that many questions. I survived that interview and received an offer from Dale. That was the only offer I received and an ENORMOUS deal for a kid who grew up in an impoverished village in a remote part of China. In the follow-up communications, I asked Dale whether he could recommend some textbooks so I could prepare early. Dale mailed a draft of the book Numerical Ocean Circulation Modeling he just finished at that time to me in China. I really appreciated his effort to get me better prepared for my graduate study.

After I started at Rutgers, Dale asked me to work on the development of SEOM. The initial goal was to run some idealized canyon simulations and compare them with lab experiments. About a year into my study, Hernan Arango, Julia Levin, and John Wilkin in the group realized it might not be a good idea to ask a graduate student to work on the model development. So, they talked to Dale about it. From what I could tell, Dale was receptive to their suggestion. He invited me into the conversation and asked me what I was really interested in. After I told him that I preferred to work with John on applying ROMS in realistic applications, Dale generously agreed to it. So I started working with John. That was the start of my academic career.

Looking back, I realize that Dale was one of the people who have dramatically changed the course of my life. Without his help, I would not have the career and life I have now.

Gordon Zhang

Looking back, I want to say a few words about the state of ocean modeling when I met Dale in 1988. There was a well-known global model from GFDL which became MOM. There were a few other models, but they weren't open source. Dale wanted to create a new model with unique numerics, suitable for regional applications. The first version of his model was all in one model.f file (before it was named SPEM). To submit it to the Cray computer at NCAR, you had to append the Job Control Language (JCL) and then work through a front-end computer, not the Cray itself. Each computer vendor had their own operating system, some Unix of some flavor, some not. Our output was in Fortran binary, which wasn't always portable from one system to another. There was a Cray with a front end on which my output would disappear, not legal files according to the front end's operating system.

Roughly one third of the code was plotting code with calls to the NCAR graphics library. More binary output.

In 1988, NCAR had a week-long workshop on how to use their supercomputer which Roberta Young and I attended. Every time we had a short break, we'd run to Dale's office to talk about some SPEM puzzle we were working on. It was a fun but exhausting trip!

The supercomputer world was changing so fast that NCAR had to host another such workshop two years later, introducing Unicos, Cray's version of Unix.

Yes, Dale was unflappable through it all. We'd be running on a single core of a computer with less memory than your phone. Dale's genius was in coming up with problems that could be solved in such a limited environment. I'm sure he could have been happy running test cases and process studies forever - it was the funding agencies who pushed us to realistic North Atlantic and Arctic domains.

In addition to his many scientific talents, Dale was a long-time supporter of interdisciplinary modeling and a generous friend to young scientists, connecting them to the wider community of modelers - from SPEM to SCRUM to ROMS, a steady advocate for this expanding community. I especially enjoyed traveling with him to Japan in 2000 for his first PICES meeting. A remarkable, insightful, and patient man who will be sorely missed!

It is truly amazing to reflect on the state of regional ocean modeling when Dale came on the scene…the state-of-the-art was solving the barotropic vorticity equation in an open boundary domain with some of Dale’s beloved test cases (Haidvogel et al., 1980). His quest for demonstrable accuracy in simulations was legendary, and the range of computational tools that the field now enjoys has been in many ways inspired by his intellectual leadership. Such models have been applied with rigor to important problems in ocean hydrodynamics as well as interdisciplinary issues. Dale’s commitment to the latter was typified by his service as Chair of the U.S. Global Ocean Ecosystem Dynamics (GLOBEC) program’s Scientific Steering Committee, a role in which he gave selflessly for the benefit of the field. Dale was an amazing intellect, mentor, colleague, and leader—and a joy to be around, both personally and professionally. I will miss him deeply.

Dennis McGillicuddy


D.B Haidvogel, A.R Robinson, E.E Schulman, 1980. The accuracy, efficiency, and stability of three numerical models with application to open ocean problems. Journal of Computational Physics 34(1), 1-53.

It was around 2008 that Dale visited NOAA Coast Survey Development Laboratory in Silver Spring, MD with the Rutgers Ocean Modeling Group (Hernan Arango, Kate Hedstrom and John Wilkin) to deliver a 3-day ROMS training workshop. The workshop was the beginning of ROMS modeling efforts within NOAA, which now has ROMS based forecast systems in Tampa Bay, Chesapeake Bay, Delaware Bay, Cook Inlet, and the Gulf of Maine. Dale guided the workshop and directed discussions on open boundary conditions, variable bottom roughness formulations, and operational support considerations. His easy going, yet thorough, and direct approach enabled us to learn and further apply ROMS to these distinct coastal estuarine systems. His encouragement and mentoring spirit will always be with us as we undertake new and challenging coastal and estuarine modeling issues………Dick Schmalz, NOAA, retired 2013

Mohamed Iskandarani asked me to post this:

I started working with Dale in the early 90’s, he was just setting up shop at Rutgers with Kate Hedstrom when I joined his research group to develop an ocean model based on spectral elements. I had known of Dale from his articles on spectral methods when I was trying to develop my own spectral solver. Dale’s modeling group was expanding quickly back then with Aike, Julia, Tony, Hernan, Laura, John, Enrique, Gabe joining. I remember this period very fondly for its excitement and the great work atmosphere and collaborations. The research group was intellectually stimulating, very friendly and forward looking and this tone was largely set by Dale’s open, generous, and warm personality. I learned a lot of him and I regret not keeping in touch all these years.

On behave of Bernard Jean Barnier:

I first met dale during the first summer school on space oceanography organized by CNES in Grasse in 1985. I was still a Ph.D. student at FSU, and Dale was teaching numerical modeling techniques applied to oceanography. Great class, and no surprise that I decided to work with SPEM after that.

Dale's impact on our community has been really important, certainly because of his remarkable achievements, although he was always humble in front of the challenges of our science also, and very important for me, because of the personal relationship he developed with young scientists, always attentive and showing them consideration. Dale really helped me to build up confidence in myself, and I guess it has been the same for many of his collaborators.

I remember every one of my visits at Dale as great professional and personal moments, moments that I shared with Aike, whom I am also missing greatly.

Très amicalement

Bernard

Sad news...

I will always remember Dale as one of the pioneers of coastal ocean modeling, a great mentor and a good listener. He was part of a great epoch in physical oceanography, and the great leader.

On behalf of Ichiro Fukumori ...

I first met Dale in the early ‘90s. Friendly and kind, and generous with his time, Dale introduced me to ocean modeling for exploring ways to conduct ocean data assimilation. The model was SPEM, which I found to be elegantly written in a manner one could readily follow, allowing me to adapt it with ease for our study. Although I chose an application laughably coarse (a 500-km resolution North Atlantic!), Dale patiently guided me, giving me insight into the workings of GCMs that has benefited me ever since. I am grateful for the many inspirations that Dale has given me, and I will sorely miss him.

The Boulder "Daily Camera" newspaper published this obituary online today ...
https://www.legacy.com/us/obituaries/da ... d=51548008

Posted on behalf of Glenn Flierl, MIT

Dale and I talked during a time when we were both visiting NCAR; it was great seeing someone bringing more advanced numerical methods to ocean models, albeit idealized, that began to represent the eddies. This expertise in both the physics and the numerics carried through into his later coastal studies. But I'd also like to highlight his important contribution to understanding eddy dynamics in the Bretherton and Haidvogel paper, one I still refer students to.

But probably the most important influence on Dale was our introducing him to Judy when she visited us in Boulder.

Bretherton, F.P. and Haidvogel, D.B., 1976. Two-dimensional turbulence above topography. Journal of Fluid Mechanics, 78(1), pp.129-154.