My PhD supervisor died in June. I’d met with him only days earlier so that he could painstakingly revise my manuscript, giving me several hours of his precious time. He had a way of asking the very questions I didn’t want to answer, highlighting the limitations of my work that I’d been trying to hide or skim over. “You need evidence,” he’d claim, jabbing a forefinger at one of my many ‘unsubstantiated assertions’. He hated those. But he loved a good reference — although not too many for each assertion, mind you.
My supervisor, Justin Francis Costelloe, a geologist and ecohydrologist at the University of Melbourne in Australia, researched arid-zone hydrology for almost 20 years and published more than 80 peer-reviewed papers. He was eminent in his field. But for me, his real impact was in his role as a mentor to students.
Costelloe was big on time management and planning. “Is it feasible, and what is your timeline?” he’d say when I proposed something new. How I loathed preparing timelines. For the first two years of my PhD, in which I am researching groundwater management, I sloppily made them only to appease him. Now, in my third and final year, I make weekly timelines and can barely function without them. As with all great supervisors, accountability was one of Costelloe’s strong points. He wanted to make sure that I was doing what I said I was doing, to make sure that I was working.
His calm guidance kept my studies grounded. When my research direction felt like a Picasso painted during the cubist period, he told me to do something I cared about, and to trust that a research question would emerge. His instruction was logical and sequential: don’t do too much; use this programme; start here. He lit a path through the fog.
Restructuring articles was also a forte of his. I would present a study like an unshuffled deck of cards, and he would skilfully re-arrange it, putting paragraphs into a logical sequence, transforming the paper into a royal flush.
And he asked why. Always, he was asking me to rationalize things, to simplify, to generalize and explain. “What exactly is your point?” he’d say. But he remained patient, as if it weren’t the one-hundredth time he’d asked the same question of me — not to mention of all the students who came before me.
During his career, he’d drilled wells, smashed rocks and tromped through burning deserts. So, he made sure I remembered that models can only approximate the infinite complexity of natural systems. “Make sure what the model tells you makes logical sense,” he’d say. He saw the complete jigsaw, not just the disconnected pieces.
He’d send me articles that he thought I’d be interested in, and encourage me to attend conferences that would grow my professional network. Mostly, however, he would hassle me about the water budget of my groundwater model. “What are the fluxes?” “What is the model doing?” “Does it make logical sense?” I didn’t want to listen, because I was afraid of the equations and code that underpin groundwater models. But when I finally took his advice and opened the Pandora’s box of how models really work, my knowledge expanded like a rising loaf.
That’s what great mentors do: expand minds.
I wish I had told him the many ways in which he was a truly great PhD supervisor. He cared about his students. He demanded rigorous science, and led by example. He made sure I was accountable for my time and research direction. He provided guidance and direction, but did not wrap me in cotton wool. He saw the big picture of my PhD: start, middle and (soon, I hope) completion. He looked out for my future career.
His office was littered with rocks, field equipment in various states of disrepair and photos of his beautiful family. But his door was always open. He was my supervisor, yet he treated me like a colleague, valued my opinion and then gently told me why, at times, it was misguided, showing me a better way instead. But mostly he was just a great person, and I feel profoundly grateful to have been his student.
A PhD is hard. But a good supervisor makes it much easier.
Nature 562, 297 (2018)