Abstract

Natural variability in all aspects of the Earth system – including the climate system and ecosystems – presents a formidable challenge to the detection and quantification of change forced by industrial activities. Error in detection can disrupt concerted efforts to respond to the challenges of Earth system change, whereas statistically robust quantification informs our understanding of underlying mechanisms of change and helps to improve predictions into the future. The rate of observed climate change results from the superposition of natural and anthropogenic forcing and its robust estimation needs to objectively detect the timing of changes in the trend as well as their magnitude. Quantifying the rate of climate change is further challenged by “memory”, i.e. red noise, within the climate system, which can lead to overconfidence in the detected change. In this talk, I describe a new approach to separate unsteady long-term change from red noise, which clarifies a key point in the scientific debate related to the recent “hiatus” in warming.

Marine ecosystems may also respond to climate change, but detecting these impacts is hindered by the long timescale of memory within the ocean such that time-series analysis of satellite data is still inconclusive as to the sign of change in ocean chlorophyll. Here I show how utilizing both temporal and spatial dependency in the available data reveals the full uncertainty in chlorophyll trends and highlights regions undergoing significant change. More generally, the targeted development of specific statistical techniques is required to process and make full use of the rapidly growing store of Earth system data from models and advancing observational platforms. Space-time modeling in particular holds promise to improve detection of climate and environmental change in the era of big data.