CNH on Water Decisions
CNH: Understanding the Effects of Water-Use Decisions in a Dynamic Environment
This project seeks to understand, represent and simulate the coupled dynamics between social values and perceptions of water resources and hydrologic system processes. The project outcomes will assist water managers, and society broadly, to make more effective decisions concerning water availability and use under expected future conditions. There are two key components to this work: Natural Language Processing and Computational Modeling. This work is funded through NSF award BCS-1114851 (Dynamics of Coupled Natural and Human Systems).
- Water institutions network: water management exhibits both a physical infrastructure and a social infrastructure. These together comprise a network, and this network will have a particular structure. The fragmentation of the physical infrastructure, the institutions that control that infrastructure and the collective perceptions of the consumers that interact with that infrastructure potentially combine to interfere with effective management policies. We have demonstrated that natural language processing and network analyses can capture significant differences between water systems in four assessed areas (Flagstaff, AZ, Tucson, AZ, Las Vegas, NV, and the Grand Valley, CO). These differences reflect the different network structures that exist in the management strategies in these areas, from centralized to more fragmented systems.
- Murphy, J.T.; Ozik, J.; Collier, N.T.; Altaweel, M.; Lammers, R.B.; Kliskey, A.; Alessa, L.; Cason, D.; Williams, P. Water Relationships in the U.S. Southwest: Characterizing Water Management Networks Using Natural Language Processing. Water 2014, 6, 1601-1641.
- The natural language processing software was released along with the paper’s publication (https://bitbucket.org/jozik/water-skimmer)
Natural Language Processing
Water consumption occurs at individual and local levels but impacts regional and national scales. We employ computational modeling to link the dynamics that drive domestic water consumption with local and state-scale management strategies and regional hydrological impacts.
The SWIM model: “Simulating Water, Individuals and Management (SWIM)” is an agent-based model (ABM) of residential water consumption. The SWIM model represents household that respond to price shocks and to conservation messages, and whose domestic consumption is shaped by communication of norms within social networks. Seasonal and attested climate variations are controlled for, allowing investigation about the role of price vs. conservation. The SWIM model is calibrated to capture the attested downturn in consumption in Tucson, Arizona, that began around 2002 and has continued since. The model defined a balanced range of values for price sensitivity vs. conservation value transmission that could produce the observed trajectory.
The WBM and ABM Integrated Model: The UNH Water Balance/Water Transport Model (WBM) is a process-based macro-scale hydrological model with the capability to simulate natural system as well as human changes to the hydrological cycle including irrigation with multi-cropping, dams and reservoirs and inter-basin transfers of water. The WBM is integrated with the SWIM model and with an ABM of intermediate water managers. Joining these models allows: the WBM to be improved with better representation of how human decisions impact regional water; the ABM to incorporate physical hydrology across the region and among interacting cities; tracing impacts of local decisions to regional scales; examination of management network structure and its impact under regional water stress.
Initial runs have linked the two models successfully and modeled water as it is transported from the Colorado River via the Central Arizona Project to the Phoenix and Tucson Metropolitan areas. Ongoing work will use this to test regional water stress scenarios.