by Kelsey Warren 

I grew up in a drought. Parched, fading knee-high grasses, their color leached by the dry days, capped the hillsides of the central California coast. Dozens and dozens of days without a single drop of rain – this is a concept that feels close to home. So, when I joined the Rapid Climate Assessment Program (RCAP) team that was ‘Tracking Consecutive Dry Days as a Drought Precursor,’ I felt as though I would be studying a concept already familiar to me. In some ways, I was. But it’s safe to say that I had never approached my childhood drought using the R programming language. 

I spent the summer of 2026 as a graduate research assistant with the North Central Climate Adaptation Science Center (NC CASC) and this RCAP team, developing a proof of concept for a new approach to an existing drought metric: Consecutive Dry Days (CDD), a way to quantify the occurrence of dry spells. Not only was this my first experience creating a proof of concept, but it was also my first time working within a collaborative network spanning federal and state agencies. Throughout the process, I appreciated how my supervisors brought together multiple voices and perspectives as we worked through how to design the drought metric. My progression throughout the position was shaped by insights from the scientific literature, coding challenges with rewarding solutions, and a persistent feeling that I was working in a space that I have long sought: collaboratively engaging in actionable science with tangible applications. 

I began the project with a preliminary review of the literature on commonly used drought metrics. I then outlined a more comprehensive literature review for a document that would accompany the code for the new metric, building on previous work by my supervisor, Michael Downey, at the Montana Department of Natural Resources & Conservation (DNRC). After a fruitful meeting with a collaborator at the Montana Climate Office, I began conceptualizing and quantifying periods of dryness, building from my childhood intuition about drought, but this time using daily gridMET precipitation data and a statistical approach to standardizing the occurrence of dry spells. Developing the R scripts to compute this new metric, initially for the Northern Great Plains, involved a classic cycle of software crashes, consulting various online resources to troubleshoot unfamiliar errors, and satisfying “aha!” moments when each successive chunk finally ran successfully. 

As I started generating maps of dry days and normalized CDD values, the vibrant hues of R color scales splashed across those four states. Seeing these vivid shades emerge brought the satisfaction of seeing my efforts take shape, even as the fiery reds and rich purples indicated long stretches of dryness on landscapes awaiting rain. As I progressively fine-tuned the maps with feedback from my team, the process of iteration itself began to feel familiar. But this time, unlike the tawny rolling hills paling month after month, drought was marked by brighter, more intense colors the longer a landscape went dry. Through these maps, I began to understand the value of this CDD metric. The louder colors of longer dry spells called out for attention. The quantification of dry periods relative to the climatological past of each grid cell didn’t just denote low precipitation, it highlighted the dry spells themselves.

I am optimistic about the utility of this dry days metric for drought monitoring in the U.S. Conversations in this collaborative environment gave me not only ideas for how to classify and normalize dry days data, but also a sense of the existing interest in this metric within the drought monitoring community. I hope that the CDD metric will hold value not only for drought monitoring efforts, but also for water users impacted by dry spells, including farmers, land managers, conservationists, and wildfire professionals. This metric is also particularly relevant to me as a PhD student: my research looks at snowpack-driven ecosystem growth and drying processes preceding wildfires, and I would be excited to explore how CDD could be applied to snow drought in my own work. 

I’d like to express my sincere appreciation for my NC CASC RCAP team at NOAA and the Montana DNRC, as well as Heather Yocum and other graduate research assistants for their continued encouragement and support throughout this position.

About the author: Kelsey Warren is a PhD student in the Geography Department and Institute of Arctic and Alpine Research (INSTAAR) at CU Boulder. There, she studies the pre-ignition ecosystem pathways linking snow drought to wildfire. Prior to beginning her PhD studies, Kelsey earned a Master of Environmental Science and Management at the UCSB Bren School where she studied climate science and forest conservation, and a B.S. in Biology at UCLA. In her free time, you can find her trail running, surfing, or learning how to knit.