Colorado mountains
From Long-Term Data to Understanding: Toward a Predictive Ecology
2015 LTER ASM Estes Park, CO - August 30 - September 2, 2015

Responses of DOC quantity and quality to water flow and rainfall events in Tempe Town Lake: A time-series modelling approach

Printer-friendly versionPrinter-friendly version

Poster Number: 
Presenter/Primary Author: 
Hilairy Hartnett
Maria van Schaijik
Monica Palta
Albert Ruhi Vidal
Nancy Grimm

Tempe Town Lake (TTL) is a constructed lake that occupies the otherwise dry riverbed of the Salt River in Tempe, AZ. It provides the ecosystem services of flood mitigation, recreation, and aesthetic value to the region and represents a unique test-bed for studying urban aquatic biogeochemistry. The lake receives inputs of dissolved organic carbon (DOC) from rainfall, storm flows, and from upstream river discharge. Complex patterns in DOC concentration and composition suggest that carbon cycling in the lake responds to both meteorological/climatological events and anthropogenic activity. Here we present a time-series modeling approach to evaluating drivers of change in DOC concentrations, and assess the effects of these drivers on autotrophic and heterotrophic processes in the lake. First, we focused on a 3-year recent period for which high-frequency data was available. We found that flow into the lake and rainfall during the precedent week had positive influences on DOC concentration. We also investigated the possibility that the dynamics of DOC, as they relate to rainfall and flow, change once the amount of rainfall in a week surpasses a threshold. Our next step will be to examine the effects of flow and rainfall on DOC composition. We then focused on the long-term (10-year) time series, analyzing seasonal patterns in dissolved O2 concentration, and found the lake is supersaturated with respect to O2 over 70x% of the time-series. This strongly suggests the lake is autotrophic much of the time. Ongoing work involves using the ‘MARSS’ R package and extrapolating findings for the last three years to the past 10 years, for which we have monthly data, to test whether long term dynamics in the lake have changed over time. By comparing the model to the ‘observed’ past, we will be able to test the bias and precision of our model, and potentially fill gaps in periods when no DOC concentration data was available.