Sue E. Waldman
University of Oregon
Oregon State University
Columbia Plateau Conservation Research Center US Department of Agriculture
When I left college I had no idea what a mathematician did, so I began my career teaching high-school mathematics, then college mathematics. Later, when home with our small children, I was hired part time by Agricultural Research Service to do "some programming" for a few months. Twenty plus years later, I'm still at the Research Center in Pendleton, Oregon. My job has changed over the years. Originally I wrote programs - statistical analysis, curve fitting or data manipulation and storage - for the entire staff. Today, commercial software is available for many of these tasks which allowed me to return to my first love, mathematics. I am part of a team that describes plant and environmental processes mathematically or "models" them as organized patterns of interrelated theories stated as a series of equations. Eventually I program these equations into computer code to simulate environment and growth. We use such mathematical models to communicate our knowledge of processes and interactions between the physical and biological system of the crop and its environment.
A model provides an answer to a problem or question. Once the question is outlined, we read other scientists' research to become familiar with their published theories; we have "brainstorming" sessions with scientists in various fields of expertise to explore the many aspects of the problem; then many hours are spent making observations and gathering data - weather information, soil information, number and sizes of leaves, plant weight, number of seeds on a head, number of roots, etcetera. Our dataset is first analyzed graphically - the dataset is plotted using both the recognized and original theories to decide whether there are relationships with known variables such as temperature, sunshine or rainfall. I transfer these relationships of the observed data to series of equations which is called a model. These mathematical models ultimately provide cause-and-effect descriptions of crop growth and response to the environment. When we build our models using known independent variables, the output from one model can be the input for another. Finally, we bring together all of the processes and transfer them into computer code. Thus, these models become a simulation of a natural process. With a crop growth simulation, a crop can be "grown" in a computer using the current weather information. These simulations can help farmers decide, for example, whether it is economically advantageous to add fertilizer, irrigate the crop, or spray for an insect or disease.
What happens in the field can be observed, but we don't always know why it happened. It is the task of the modeler to determine a cause-and-effect relationship. Many different methods may be tried before we discover what works, but in research we must be willing to make mistakes and recognize errors. I enjoy being a mathematician/modeler/programmer in agricultural research. It requires creativity and problem-solving skills; there are all kinds of things to be discovered and endless challenging questions waiting for solutions. In applying mathematics to the "real world", we observe and measure the world in such detail that we are able to see its simplicity and then communicate its process mathematically.