Modeling Bioeconomic Risk States in Guadalupe Valley, Mexico

Bioeconomic risk analysis and modeling with focus on groundwater resources was performed in the Guadalupe Valley, a region with growing wine industry located in a semi-arid zone in Northwest Mexico. Bioeconomy is an instance of a system where socioeconomic activities and biophysical processes are tightly coupled. Identifying the critical states of such a system requires understanding of the human-modified water cycle with specific inclusion of the mutual feedback between human and hydrologic components.

A systems perspective was used to design an integrated conceptual model that captures key components related to the coupled socio-hydrological system. The conceptual model proposed in this study combines the DPSIR environmental framework with the SPRC risk framework. This enabled a holistic, qualitative description of system states and feedbacks and guided the design and input data selection for the modeling stage.

Our study complements more physically-based approaches in using data-driven methods. We use Support Vector Machine (SVM) algorithms to integrate multivariate and multitemporal data on human behavior, socioeconomy, climate, land and vegetation to discover their interactions with hydrologic processes. In utilizing both radar and optical remote sensing data in our analysis we establish a link from free and open data monitoring systems with improved spatial and temporal resolution to phenomena particularly linked to human water use, infrastructure and land development. The water budget components were determined (PANTA RHEI software) and groundwater flow, balancing supply and storage was simulated using a high-resolution 3d numerical groundwater flow model (FEFLOW). All those were used as proxies for the model variables identified in the conceptualization.

A severe hydrological drought in 2014 served as starting point for data collection. Even though precipitation recovered during the next 3 years, the deterioration of the bioeconomic system continued. According to official records from national agencies of water and agriculture, the Valley showed a groundwater deficit of 18.2 hm³ and a 25% loss of agricultural production in 2017. Therefore degradation of groundwater resources, soil productivity and negative impacts on society are key critical states that were formalized and quantified. The first SVM models allowed identifying and structure components that have greater contribution to the risk states of the Guadalupe Valley bioeconomic system.

Both the conceptual and the data-driven SVM model are useful tools to understand the mechanisms of risk generation and to promote the evaluation of regional water management strategies.

This research is part of the bi-national ECOAQUA project funded by BMBF, Germany and CONACYT, Mexico.