Keywords

Agent-based modeling; Characterization and Parameterization Framework (CAP); Green Infrastructure; Innovation diffusion; Stormwater management.

Start Date

5-7-2022 12:00 PM

End Date

8-7-2022 10:00 AM

Abstract

Incorporating social-science theories to represent human decision-making in data-driven models is required, particularly when modeling the underlying social determinants of adopting pro-environmental innovations. Our agent-based model (ABM) investigates households' adoption of stormwater green infrastructure (GI) - e.g., rain gardens - on private residential yards in Kitchener/Waterloo, Ontario. We propose combining Rogers's diffusion of innovation theory (DOI) (2003) with Smajgl Barreteau (2017) characterization and parameterization framework (CAP) to develop an empirical ABM, DRAIIN. At the micro-level, we use a municipal survey (Aquafor Beech Ltd., Freeman Associates 2015) to parameterize agents' attributes (e.g., ownership status) and behavioral data inputs (e.g., yard maintenance willingness). Additionally, a local survey (Defields, 2013) feeds into a multinomial logit model to calibrate agents' adoption decisions. Since the sample data is representative of the case study population, we apply cloning to upscale the sampled agents proportionally. Consistent with comparable case studies (Nassauer, Wang, Dayrell, 2009), both the municipal and our local surveys suggest that peer influences primarily influence households' landscape decisions. Similar market dynamics are observed in the solar panels market (Bollinger, Gillingham, 2012; Rode Weber, 2016). Thus, we postulate that DRAIIN follows the generalized S-curve for innovation diffusion that was reported in solar panels' cumulative market growth (Islam, 2014; Palmer, Sorda, Madlener, 2015). At the macro-level, we seek qualitative validation by examining whether DRAIIN generates the diffusion S-curve (see Rand Rust, 2011). Following the CAP framework, we incorporate expert knowledge from city representatives, stormwater consultants, and a local not-for-profit to co-validate DRAIIN's structure and outputs. Integrating the DOI and the CAP framework is a context-independent approach that is potentially applicable to other cases of innovation-diffusion ABMs beyond DRAIIN. Further, we suggest that incorporating empirical data into Rogers's innovation-decision stages should enhance the predictive power of innovation-diffusion models, which inherently lack post-diffusion validation data.

Stream and Session

false

Share

COinS
 
Jul 5th, 12:00 PM Jul 8th, 10:00 AM

Employing the Characterization and Parameterization framework (CAP) to develop an empirical innovation-diffusion agent-based model of green infrastructure adoption on private residential yards.

Incorporating social-science theories to represent human decision-making in data-driven models is required, particularly when modeling the underlying social determinants of adopting pro-environmental innovations. Our agent-based model (ABM) investigates households' adoption of stormwater green infrastructure (GI) - e.g., rain gardens - on private residential yards in Kitchener/Waterloo, Ontario. We propose combining Rogers's diffusion of innovation theory (DOI) (2003) with Smajgl Barreteau (2017) characterization and parameterization framework (CAP) to develop an empirical ABM, DRAIIN. At the micro-level, we use a municipal survey (Aquafor Beech Ltd., Freeman Associates 2015) to parameterize agents' attributes (e.g., ownership status) and behavioral data inputs (e.g., yard maintenance willingness). Additionally, a local survey (Defields, 2013) feeds into a multinomial logit model to calibrate agents' adoption decisions. Since the sample data is representative of the case study population, we apply cloning to upscale the sampled agents proportionally. Consistent with comparable case studies (Nassauer, Wang, Dayrell, 2009), both the municipal and our local surveys suggest that peer influences primarily influence households' landscape decisions. Similar market dynamics are observed in the solar panels market (Bollinger, Gillingham, 2012; Rode Weber, 2016). Thus, we postulate that DRAIIN follows the generalized S-curve for innovation diffusion that was reported in solar panels' cumulative market growth (Islam, 2014; Palmer, Sorda, Madlener, 2015). At the macro-level, we seek qualitative validation by examining whether DRAIIN generates the diffusion S-curve (see Rand Rust, 2011). Following the CAP framework, we incorporate expert knowledge from city representatives, stormwater consultants, and a local not-for-profit to co-validate DRAIIN's structure and outputs. Integrating the DOI and the CAP framework is a context-independent approach that is potentially applicable to other cases of innovation-diffusion ABMs beyond DRAIIN. Further, we suggest that incorporating empirical data into Rogers's innovation-decision stages should enhance the predictive power of innovation-diffusion models, which inherently lack post-diffusion validation data.