Keywords
modelling methodology; data integration; communication
Start Date
5-7-2022 12:00 PM
End Date
8-7-2022 9:59 AM
Abstract
Models and data (and those who generate them) have an important but sometimes uneasy relationship. Effective integration of information from models and data is the “holy grail” for decision-makers and those seeking to understand environmental systems. Data provide snapshots of what is happening in the system, which is essential for thorough model calibration, validation, and testing. In contrast, models can explore the underlying processes driving the observed behaviour. Thus, models and data provide complementary information about a system. In cases where model predictions do not agree with observations, resolving any discrepancies is essential for model-data integration and can be a valuable source of insight. The resolution of such discrepancies often requires engagement between the modellers and those who collected the data. However, this relationship may not be comfortable since those who build and run models, and those who collect data can have different viewpoints, constraints, and priorities, as well as different purposes for models or data. This talk addresses both of these issues by providing 1. A systematic framework for dealing with model-data discrepancies, illustrated by case studies; and 2. An action plan to improve relationships between those who model and those who measure. These results will be presented in traditional academic form and then in a light-hearted animation, in which a modeller and data-collector are sent to marriage counselling to resolve their differences. The take-home message is that integrating model and data insights is best done if there are strong relationships between those building and running models and those collecting data, and good relationships take effort.
Model-data marriage counselling: resolving the age-old conflict between models and data (and between those who model and those who measure)
Models and data (and those who generate them) have an important but sometimes uneasy relationship. Effective integration of information from models and data is the “holy grail” for decision-makers and those seeking to understand environmental systems. Data provide snapshots of what is happening in the system, which is essential for thorough model calibration, validation, and testing. In contrast, models can explore the underlying processes driving the observed behaviour. Thus, models and data provide complementary information about a system. In cases where model predictions do not agree with observations, resolving any discrepancies is essential for model-data integration and can be a valuable source of insight. The resolution of such discrepancies often requires engagement between the modellers and those who collected the data. However, this relationship may not be comfortable since those who build and run models, and those who collect data can have different viewpoints, constraints, and priorities, as well as different purposes for models or data. This talk addresses both of these issues by providing 1. A systematic framework for dealing with model-data discrepancies, illustrated by case studies; and 2. An action plan to improve relationships between those who model and those who measure. These results will be presented in traditional academic form and then in a light-hearted animation, in which a modeller and data-collector are sent to marriage counselling to resolve their differences. The take-home message is that integrating model and data insights is best done if there are strong relationships between those building and running models and those collecting data, and good relationships take effort.
Stream and Session
false