Abstract
There are many parameters involved in the bloom patterns of cyanobacteria in Lake Torment. This includes chemical components, viz., TP, PO4-P, NO3-N, NH4-N, and iron etc. However, if there exists multicollinearity in those involving in the CHAB patterns with a quasi-linear dependence in which two or more than two predictor variables are highly correlated to each other, we often cannot control the predictors/drivers affecting on the CHAB appearance. We expect therefore to evaluate the multicollinearity of key parameters leading to the significant contribution of our CHAB predicting mathematical framework, and based on it, multi-regression models have been suggested to predict the CHAB patterns.
Similar content being viewed by others
Availability of Data and Materials
The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.
References
Hushchyna, K., Vo, Q. B., & Nguyen-Quang, T. (2020). The application of TRINDEX to predict Harmful Algal Blooms in Lake Torment (NS, Canada). Environmental Problems., 5(3), 156–163.
Marty, J. (2018). Kings County Lake monitoring program 2018 season. Report. Municipality of the County of Kings.
Bi, J. (2012). A Review of the Statistical Methods for Determination of Relative Importance of Correlated Predictors and Identification of Drivers of Consumer Liking. Journal of Sensory Studies, 27(2), 87–101. https://doi.org/10.1111/j.1745-459X.2012.00370.x
Achen, C. H. (1982). Interpreting and Using Regression. Sage. Beverly Hills. CA.
Stine, R. A. (1995). Graphical Interpretation of Variance Inflation Factors. The American Statistician, 49(1), pp. 53–56. http://www.jstor.org/stable/2684812?origin=JSTOR-pdf
Lindeman, R. H., Merenda, P. F., & Gold, R. Z. (1980). Introduction to bivariate and multivariate analysis. Scott Foresman. Retrieved November 5 2022, from http://catalog.hathitrust.org/api/volumes/oclc/5310754.html
Grömping, U. (2006). Relative importance for linear regression in R: the package relaimpo. Journal of statistical software, 17, 1-27. https://doi.org/10.18637/jss.v017.i01
Kelley, K., & Bolin, J. H. (2013). Multiple Regression. In Handbook of quantitative methods for educational research, pp. 69–101. Brill Sense.
Hair, J. F, Anderson, R. E, Babin, B. J, & Black, W. C. (2010). Multivariate data analysis: A Global Perspective. 7th ed. Upper Saddle River (N.J.): Pearson education.
Acknowledgements
TNQ is thankful for the NSERC Discovery Grant RGPIN 03906. The authors would like to thank Lake Torment residents, especially Alexander’s family, for their support.
Funding
NSERC Discovery Grant RGPIN 03906.
Author information
Authors and Affiliations
Contributions
Conceptualization: Tri Nguyen-Quang and Kateryna Hushchyna; methodology: all; software: Tri Nguyen-Quang, Qurat Ul An Sabir, and Kayla Mclellan; validation: Qurat Ul An Sabir and Kateryna Hushchyna; formal analysis: Tri Nguyen-Quang, Qurat Ul An Sabir, and Kateryna Hushchyna; investigation: Kayla Mclellan, and Kateryna Hushchyna; resources: Tri Nguyen-Quang, Qurat Ul An Sabir, and Kateryna Hushchyna; data curation: Kateryna Hushchyna and Kayla Mclellan; writing—original draft preparation: Kateryna Hushchyna and Tri Nguyen-Quang; writing, review, and editing: Tri Nguyen-Quang and Qurat Ul An Sabir; visualization: Kateryna Hushchyna and Qurat Ul An Sabir; supervision: Tri Nguyen-Quang; project administration: Tri Nguyen-Quangand Qurat Ul An Sabir.
Corresponding author
Ethics declarations
Ethics Approval
Not applicable.
Consent to Participate
All authors gave consent to participate in a study, and to have their data published in this article.
Consent to Publication
All authors have given consent to publication.
Competing Interests
The authors declare no competing interests.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Annex 1: Validation of the Regression Models
Real Data | Validation data | |||||
---|---|---|---|---|---|---|
Day (locations) | Real PC | Real CHL-A | Model 1 (PC) | Model 2 (CHL-A) | Relative error for PC (%) | Relative error for Chl-a (%) |
July 19, 2020 (T8) | 1 | 0.644 | 0.985 | 0.606 | 1.475 | 6.271 |
July 21, 2020 (T8) | 1 | 0.655 | 1.003 | 0.601 | 0.264 | 8.943 |
July 22, 2020 (T8) | 1 | 0.647 | 0.995 | 0.595 | 0.494 | 8.721 |
July 25, 2020 (Toutlet) | 1 | 0.656 | 1.008 | 0.593 | 0.808 | 10.645 |
Oct 1, 2020 (T8) | 1 | 0.656 | 1.008 | 0.593 | 0.808 | 10.645 |
Oct 2, 2020 (T8) | 0.016 | 0.003 | 0.017 | 0.019 | 7.719 | 5.111 |
Nov 8, 2020 (T8) | 0.99 | 0.675 | 1.027 | 0.6 | 3.625 | 12.588 |
Nov 8, 2020 (T4) | 1 | 0.647 | 0.995 | 0.595 | 0.494 | 8.721 |
Nov 8, 2020 (T8) | 1 | 0.647 | 0.995 | 0.595 | 0.494 | 8.721 |
Nov 8, 2020 (T4) | 1 | 0.647 | 0.996 | 0.593 | 0.374 | 9.158 |
Nov 8, 2020 (Toutlet) | 1 | 0.647 | 0.995 | 0.596 | 0.519 | 8.63 |
Nov 8, 2020 (T9) | 1 | 0.659 | 1.011 | 0.594 | 1.121 | 10.867 |
Nov 8, 2020 (T8) | 1 | 0.648 | 0.991 | 0.605 | 0.89 | 7.108 |
Nov 8, 2020 (T9) | 1 | 0.648 | 0.997 | 0.593 | 0.264 | 9.296 |
Aug 2, 2021 (T8) | 1 | 0.648 | 0.997 | 0.594 | 0.335 | 9.033 |
Aug 6, 2021 (T8) | 0.033 | 0.028 | 0.308 | 0.025 | 8.178 | 9.919 |
Aug 7, 2021 (T8) | 1 | 0.648 | 0.991 | 0.605 | 0.863 | 7.141 |
Aug 10, 2021 (T8) | 0.033 | 0.659 | 0.033 | 0.594 | 1.797 | 10.86 |
Aug 13, 2021 (T8) | 0.248 | 0.174 | 0.246 | 0.161 | 0.519 | 7.9 |
Aug 17, 2021 (T8) | 0.248 | 0.174 | 0.246 | 0.161 | 0.519 | 7.9 |
Sept 12, 2021 (T8) | 0.248 | 0.174 | 0.246 | 0.161 | 0.519 | 7.9 |
Sept 15, 2021 (T8) | 0.248 | 0.174 | 0.246 | 0.161 | 0.519 | 7.9 |
Sept 15, 2021 (Toutlet) | 0.248 | 0.174 | 0.247 | 0.161 | 0.202 | 8.118 |
Oct 14, 2021 (T8) | 0.248 | 0.174 | 0.246 | 0.161 | 0.519 | 7.9 |
Oct 14, 2021 (T9) | 0.253 | 0.499 | 0.248 | 0.501 | 2.173 | 0.446 |
Oct 14, 2021 (Toutlet) | 1 | 0.605 | 0.934 | 0.605 | 7.106 | 0.047 |
Oct 14, 2021 (T9) | 0.245 | 0.103 | 0.257 | 0.117 | 4.57 | 12.362 |
Oct 14, 2021 (T9) | 0.833 | 0.499 | 0.795 | 0.488 | 4.783 | 2.308 |
June 3, 2022 (T9) | 0.833 | 0.499 | 0.79 | 0.498 | 5.489 | 0.106 |
June 9, 2022 (T9) 9am | 0.833 | 0.499 | 0.79 | 0.498 | 5.489 | 0.106 |
June 15, 2022 (T5) | 1 | 0.605 | 0.939 | 0.594 | 6.516 | 1.829 |
June 15, 2022 (T8) | 0.722 | 0.499 | 0.685 | 0.466 | 5.387 | 6.975 |
June 15, 2022 (T4) | 0.627 | 0.499 | 0.687 | 0.543 | 8.772 | 8.13 |
June 15, 2022 (T9) | 0.712 | 0.499 | 0.683 | 0.464 | 4.233 | 7.421 |
June 15, 2022 (T9) | 0.627 | 0.499 | 0.685 | 0.543 | 8.536 | 8.13 |
June 15, 2022 (T9) | 0.77 | 0.499 | 0.683 | 0.501 | 12.726 | 0.446 |
June 15, 2022 (Tout) | 0.833 | 0.499 | 0.79 | 0.498 | 5.489 | 0.106 |
Annex 2: VIF Calculation for Each Predictor Versus Equivalent Target
Factor | VIF value for Chl-a |
---|---|
Color | 33.275 |
TP | 30.391 |
Cond | 5.506 |
NH4_N | 3.890 |
T | 3.849 |
SiO2 | 3.081 |
PC | 2.811 |
NO3_N | 2.761 |
Iron | 2.620 |
PO4_P | 2.572 |
DO | 2.565 |
pH | 1.575 |
Factor | VIF value for PC |
TP | 40.531 |
Color | 34.435 |
Chl-a | 20.844 |
Cond | 5.525 |
NH4_N | 3.380 |
T | 3.482 |
SiO2 | 2.947 |
NO3_N | 2.691 |
Iron | 2.462 |
PO4_P | 2.454 |
DO | 2.359 |
pH | 1.684 |
Factor | VIF value for MCLR |
Chl-a | 335.400 |
TP | 285.691 |
Color | 81.137 |
PC | 52.52 |
Cond | 30.228 |
T | 16.279 |
Iron | 13.575 |
SiO2 | 9.534 |
NH4_N | 9.189 |
PO4_P | 7.925 |
pH | 3.452 |
NO3_N | 3.174 |
DO | 2.848 |
Annex 3: LMG index (index for the variance importance)
A3.1. Relative importance metrics LMG for Chl-a |
---|
PC 0.156511076 15.6% DO 0.013332912 T 0.013776778 pH 0.004292141 PO4_P 0.020448693 NH4_N 0.011028286 NO3_N 0.006099871 TP 0.318641102 31.9% Iron 0.077737935 7.77% SiO2 0.043251999. Cond 0.027754583. Color 0.299859759 29.99% |
A3.2. Relative importance metrics LMG for PC |
Chl-a 0.38744707 38.7% DO 0.03644179 T 0.02508286 pH 0.01117246 PO4_P 0.04421082 NH4_N 0.08106255 8.11% NO3_N 0.01160576 TP 0.12809728 12.8% Iron 0.04916675. SiO2 0.03057672. Cond 0.04421614. Color 0.09706639 9.71% |
A3.3. Relative importance metrics LMG for MCLR |
Chl-a 0.169253034 16.9% DO 0.085182256 8.52% T 0.047486642 pH 0.006301402 PO4_P 0.049363006 NH4_N 0.027271629 NO3_N 0.032489614 TP 0.192465775 19.2% Iron 0.061680842 6.17% SiO2 0.047992685. Cond 0.025199578. Color 0.164947148 16.5% PC 0.031347563 |
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Hushchyna, K., Sabir, Q.U.A., Mclellan, K. et al. Multicollinearity and Multi-regression Analysis for Main Drivers of Cyanobacterial Harmful Algal Bloom (CHAB) in the Lake Torment, Nova Scotia, Canada. Environ Model Assess 28, 1011–1022 (2023). https://doi.org/10.1007/s10666-023-09907-z
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10666-023-09907-z