Submitted to: American Geophysical Union
Publication Type: Abstract Only
Publication Acceptance Date: September 11, 2008
Publication Date: December 1, 2008
Citation: Keener, V.W., Jones, J.W., Bosch, D.D., Lowrance, R.R. 2008. An ENSO-Based Multivariate Wavelet-Decomposition Time Series Model for Nitrate Loads in the Little River Watershed. American Geophysical Union. Technical Abstract: The El-Nino/Southern Oscillation (ENSO) is a periodic global climate phenomenon with strong effects on the weather patterns of the southeast United States. ENSO has been shown to have predictable effects on stream flow, rainfall, crop yield, and nutrient loads in runoff. In monitoring and research efforts during the last century, ENSO indices have emerged as one of the most consistent for describing low-frequency climate variability on both global and regional scales. To better understand the relationship between Sea Surface Temperature (SST) anomalies in the equatorial Pacific Ocean and hydrology and climate in the southeast United States, we have done an analysis in the frequency domain on 30 years of SST, precipitation, flow, and nutrient load data from an agricultural coastal plain watershed in Tifton, Georgia. To specifically understand the low-frequency oscillations and inter-annual or decadal variability inherent in these hydrological time series as a non-stationary process, wavelet analysis was used. We found that the 3-7 year mode of variability known in ENSO cycles exists in the Little River Watershed's precipitation, flow, nitrate and total phosphorus load time series. SST's and both nutrient loads, stream flow and precipitation time series also demonstrated shared periodicity and high covariance and correlation in the 3-7 year period in cross and coherence wavelet analysis. This indicates that the ENSO signal could be used as a predictor for nutrient loads in the southeast United States. Reconstructed Components (RC) from the 3-7-year period of each time series were extracted and used to create an optimized local monthly multivariate time series nutrient load model based on SST’s and the most relevant hydrological variables investigated. This predictor of nutrient loads can then be compared to a process-based hydrological model simulation of the same system. This model could be useful for land and water managers in the southeast United States, as high risk months for greater than average nutrient loading could be identified and managed in advance, based on predictions of ENSO phase.