Bureau of Reclamation – Lower Colorado Regional Office
Contract to University of Colorado for Drought Study on the Colorado River
Scope of Work
Background
The Western US is reeling under a severe dry spell the last five years. While this region is generally semiarid and is no stranger to dry periods, but the recent dry spell coupled with increasing population and economic growth and, competing water demands, is having a decisive impact on the socioeconomic well being of the region. The vast networks of reservoirs that sustain and fuel development are alarmingly at their historic lows, thereby, threatening development. The dry period is also having an impact on the water quality e.g., low flows in the streams lead to increased concentration of salinity thus, degradation of water quality. Furthermore, compacts and decrees to share water resources among the Western states developed during wet periods of early 1920s and 1930s, are under stress especially, as the states confront economic growth under limited water resources.
There is increasing evidence that the yeartoyear variations in the Western U. S. hydro-climate are driven by largescale climate features (e.g., ENSO, PDO, AMO etc.). Hoerling and Kumar (2003), from climate model simulations, show tantalizing evidence that the dry spell could be a result of cooler than normal tropical Eastern Pacific and warmer than normal tropical Western Pacific (a La Nina pattern) and Indian Oceans. This long break from the El Ninoactive period of 1980s and 1990s could be a strong factor for the dry spell. Which then raises the question why now? Adding to these woes are recent indications (Regonda et al., 2003; Cayan et al., 2002, Mote, 2003) that the annual cycle in precipitation is shifting early in the Western US perhaps, caused by global climate change. In particular, the spring warmth is tending to occur early with decreased winter snow pack. This implies reduced inflow into the rivers, as a majority of them are driven by snowmelt.
Given the high stakes, two immediate questions and requirements to USBR for efficient water resources planning and management are:
1. How unusual is the current dry spell?
2. How can we simulate stream flow scenarios that are consistent with the current dry spell and other realistic conditions?
Answering the first question will place the current dry period in context; answering the second question will provide a framework that can incorporate such information in generating realistic stream flow scenarios for driving planning and management models.
Approach
This work will consist of the following steps:
1. Long records of tree ring reconstructions of stream flows on the Colorado basin (at key locations) will be compiled. Markov Chain transition probabilities (i.e. probability of transitioning into a dry state in the follow year from a wet state in the current year, etc.) will be estimated in a moving window. Thus, the temporal variability of these transition probabilities will be observed. Their links with largescale climate features (ENSO, PDO) will also be investigated. (Time period : 2.5 months)
2. The transitional probabilities will be used to generate long synthetic sequences of the state of the system (i.e. wet or dry) and consequently, the probabilities of long dry spells. Thus, the extreme nature of the current dry spell can be placed in context. Transitional probabilities from the different epochs can be used to generate different wet and dry spells. (Time period: 3 months)
3. Conditioned on the state of the system, the flow magnitudes will be generated. Here we will use the well-tested nonparametric methods. (Time period: 6.5 months)
The advantages of this approach are:
a) Long records of stream flows from tree rings can help provide insights into the transitional behavior of the stream flows and consequently, the propensity for extended dry and wet spells
b) Nonparametric stream flow generation methods can simulate flow scenarios consistent with the state of the system.
Working with Reclamation personnel, the generated flow scenarios will be used to drive decision models (i.e. RIVERWARE) to evaluate the system reliability and other key decision variables.
Contractor:
Regents of the University of Colorado
572 UCB
Boulder, Colorado 80309-0572
Contractor’s Technical Contact:
Dr. Balaji Rajagopalan
Assistant Professor and Fellow, Cooperative Institute for Research in Environmental Sciences
Dept. of Civil, Environmental and Architectural Engineering
Campus Box 428, ECOT541
University of Colorado
Boulder, CO 803090428, USA
Phone:(303)4925968(w); Fax: (303)4927317
email:
Dates: September 1, 2004 to December 31, 2005
Total Budget: $24,500