One of the most frequently-performed analysis of time series data is finding a trend. In finance, a trend tells an executive summary of the state of the economy (surmised from high quality “big data” sets), in marketing, trends tell the stories of social preferences (presumably “big data” sets previously unavailable make the stories more robust). In climate science, the global temperature trend is used as a proxy for defining what’s normal and what’s not normal “climate.” In principle, this global temperature trend is a reasonable quantity of interest, as it relates to the most fundamental thermodynamic balance in Earth’s climate: radiation. This talk will disappoint those who want to know with quantitative precision what the trend is on the global temperature, beyond just merely increasing. Climate change trends are undoubtedly important. I will show you how challenging I found trying to define a trend is, how scientifically challenging the task of defining a trend on our 1 experiment we call Earth’s Climate is, how this quest lead me to diffusion maps and universality in time series spectra, and the consideration of multi-scale analysis of time series data; and finally, to suggest that this is a problem that lends itself for tight research collaborations between geoscientists, mathematicians, statisticians, and engineers.