An approach for finding causal relations in environmental systems: with an application to understand drivers of a toxic algal bloom
Abstract Background Discovering causality in environmental systems is challenging because frequently controlled experiments or numerical simulations are difficult. Algorithms to learn directed acyclic graphs from system data are powerful, but they often result in too many possible causal structures...
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2024-01-01
|
| Series: | Ecological Processes |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s13717-023-00482-5 |
| _version_ | 1797275832891211776 |
|---|---|
| author | Benny Selle |
| author_facet | Benny Selle |
| author_sort | Benny Selle |
| collection | DOAJ |
| description | Abstract Background Discovering causality in environmental systems is challenging because frequently controlled experiments or numerical simulations are difficult. Algorithms to learn directed acyclic graphs from system data are powerful, but they often result in too many possible causal structures that cannot be properly evaluated. Results An approach to this problem proposed here is to initially restrict the system to a target variable with its two major drivers. Subsequently, testable causal structures are obtained from rules to infer directed acyclic graphs and expert knowledge. The proposed approach, which is essentially based on correlation and regression, was applied to understand drivers of a toxic algal bloom in the Odra River in summer 2022. Through this application, useful insight on the interplay between river flow and salt inputs that likely caused the algal bloom was obtained. Conclusions The Odra River example demonstrated that carefully applied correlation and regression techniques together with expert knowledge can help to discover reliable casual structures in environmental systems. |
| first_indexed | 2024-03-07T15:19:42Z |
| format | Article |
| id | doaj.art-24c783776dc644acb9a68211a01463ba |
| institution | Directory Open Access Journal |
| issn | 2192-1709 |
| language | English |
| last_indexed | 2024-03-07T15:19:42Z |
| publishDate | 2024-01-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | Ecological Processes |
| spelling | doaj.art-24c783776dc644acb9a68211a01463ba2024-03-05T17:42:21ZengSpringerOpenEcological Processes2192-17092024-01-0113111210.1186/s13717-023-00482-5An approach for finding causal relations in environmental systems: with an application to understand drivers of a toxic algal bloomBenny Selle0Berliner Hochschule Für TechnikAbstract Background Discovering causality in environmental systems is challenging because frequently controlled experiments or numerical simulations are difficult. Algorithms to learn directed acyclic graphs from system data are powerful, but they often result in too many possible causal structures that cannot be properly evaluated. Results An approach to this problem proposed here is to initially restrict the system to a target variable with its two major drivers. Subsequently, testable causal structures are obtained from rules to infer directed acyclic graphs and expert knowledge. The proposed approach, which is essentially based on correlation and regression, was applied to understand drivers of a toxic algal bloom in the Odra River in summer 2022. Through this application, useful insight on the interplay between river flow and salt inputs that likely caused the algal bloom was obtained. Conclusions The Odra River example demonstrated that carefully applied correlation and regression techniques together with expert knowledge can help to discover reliable casual structures in environmental systems.https://doi.org/10.1186/s13717-023-00482-5Causal effectCausal modelCausal diagramMediation analysisDirected acyclic graphClassification and regression trees |
| spellingShingle | Benny Selle An approach for finding causal relations in environmental systems: with an application to understand drivers of a toxic algal bloom Ecological Processes Causal effect Causal model Causal diagram Mediation analysis Directed acyclic graph Classification and regression trees |
| title | An approach for finding causal relations in environmental systems: with an application to understand drivers of a toxic algal bloom |
| title_full | An approach for finding causal relations in environmental systems: with an application to understand drivers of a toxic algal bloom |
| title_fullStr | An approach for finding causal relations in environmental systems: with an application to understand drivers of a toxic algal bloom |
| title_full_unstemmed | An approach for finding causal relations in environmental systems: with an application to understand drivers of a toxic algal bloom |
| title_short | An approach for finding causal relations in environmental systems: with an application to understand drivers of a toxic algal bloom |
| title_sort | approach for finding causal relations in environmental systems with an application to understand drivers of a toxic algal bloom |
| topic | Causal effect Causal model Causal diagram Mediation analysis Directed acyclic graph Classification and regression trees |
| url | https://doi.org/10.1186/s13717-023-00482-5 |
| work_keys_str_mv | AT bennyselle anapproachforfindingcausalrelationsinenvironmentalsystemswithanapplicationtounderstanddriversofatoxicalgalbloom AT bennyselle approachforfindingcausalrelationsinenvironmentalsystemswithanapplicationtounderstanddriversofatoxicalgalbloom |