Intensive agricultural practices can dramatically change the landscape, thereby increasing the concentrations and rates at which nutrients are delivered to aquatic ecosystems. In the United States, concerns about accelerating rates of lake eutrophication related to increases in nutrient loading require a method of quantifying ecological changes that have occurred since European settlement. Because the application of traditional quantitative total phosphorus transfer functions in paleolimnology has proven difficult in shallow, hypereutrophic lakes, we used several approaches in this study to assess ecosystem changes associated with eutrophication of 32 natural lakes in the state of Iowa, USA. In addition to traditional transfer function methods, we estimated changes in primary productivity from the flux of biogenic silica (BSi) and organic carbon accumulation rates (OC AR). Additionally, we compared pre-disturbance diatom communities to modern diatom communities, i.e. floristic change, using non-metric multi-dimensional scaling and square chord distance. OC AR and BSi fluxes increased over time and were positively correlated with the time period of agricultural intensification in the region (post-1940). Ninety-one percent of the lakes in this study showed evidence for eutrophication based on geochemical proxies, and 88 % of lakes showed major floristic change in the diatom community. Whereas geochemical indicators showed consistent increases in productivity across most lakes, floristic changes reflected more complex interactions between other environmental drivers. The magnitude of floristic change did not directly correlate to nutrient-driven increases in primary production, but was driven by ecological diatom assembly related to lake depth. Transfer functions consistently perform poorly, especially for shallow lakes, and other techniques that combine geochemistry and diatom ecology are recommended for reconstructions of eutrophication.