The Use of Oceanic Indices Variations Due to Climate Change to Predict Annual Discharge Variations in Northeastern United States

Abstract:

Climatic change affects streamflow in watersheds with winter snowpack and an annual snowmelt hydrograph. In the northeastern US, changes in streamflow are driven by both the advanced timing of snowmelt and increasing summer precipitation. Projections of climate for the region in the 21^st century is for warmer winters and wetter summers. Water planners need to understand future changes in flow metrics to determine if the current water resources are capable of fulfilling future demands or adapting to future changes in climate. The study of teleconnection patterns between oceanic indices variations and hydrologic variables may help improve the understanding of future water resources conditions in a watershed.

The purpose of this study is to evaluate the correlation between oceanic indices and discharge variations in the Merrimack Watershed. The Merrimack Watershed is the fourth largest basin in New England which drains much of New Hampshire and northeastern portions of Massachusetts, USA. Variations in sea surface temperature (SST) and sea level pressure (SLP) are defined by the Atlantic Multi-decadal Oscillation (AMO) and the North Atlantic Oscillation (NAO), respectively. We hypothesize that temporal changes in discharge are related to AMO and NAO variations since precipitation and discharge are highly correlated in the Merrimack.

The Merrimack Watershed consists of undisturbed (reference) catchments and disturbed (developed) basins with long stream gauge records (> 100 years). Developed basins provide an opportunity to evaluate the impacts of river regulation and land development on teleconnection patterns as well as changing climate. Time series of AMO and NAO indices over the past 150 years along with Merrimack annual precipitation and discharge time series have shown a 1 to 2-year watershed hydrologic memory; higher correlation between Merrimack‎ annual precipitation and discharge with AMO and NAO are observed when a 1 to 2-year lag is given to AMO and NAO indices. For instance, the mean correlation of AMO with precipitation/discharge for a zero-year lag was 0.16/0.09 and increased to 0.26/0.23 for a 1-year lag. Our study provides an insight on the lagged hydrologic response of reference catchments and developed basins to variations in oceanic indices.