Beyond First Order Kinetics: Sensitivity, Identifiability and the Limits of An Extended Quadratic Version for Petroleum Hydrocarbon Bioremediation: Local, Global and Profile-Likelihood Analysis

Abstract A companion study showed that an extended version of the first-order kinetic model, the Hindrance-Modified First-Order Kinetic Model (HM-FOKM), B(t)=1-exp⁡(-Kt+H/2 t^2 ), outperforms classical first-order kinetics on AICc in 15 of 24 treatments that exhibit genuine quadratic curvature. This study assesses the sensitivity and identifiability of this model through local and global sensitivity analysis, profile-likelihood identifiability analysis, and a new diagnostic: the time to 90 % removal (t_90). The local sensitivity functions S_Kand S_Hmaintain fixed signs throughout the domain (S_K≥0, S_H≤0), a simpler structure than that of models whose sensitivities change sign. Because Kand Hare strongly correlated (mean ∣ρ∣=0.924), independent-box Sobol sampling is invalid. A correlation-ratio ensemble reveals that K dominates early (S_1 (K)=0.98 at 1 week), while Hovertakes Kat approximately 9.8 weeks Profile-likelihood analysis confirms that Kis better constrained than H. The 95 % confidence interval for Hincludes zero in 10 of 24 treatments (41.7 %), so classical first-order kinetics cannot be excluded even when HM-FOKM is preferred by AICc. The correlation ∣ρ(Kⓜ,H)∣ranges from 0.890 to 0.982, consistent with a general collinearity pattern in two-parameter saturating kinetic models. The t_90diagnostic exists only when K^2≥2H⋅ln⁡10(equivalently B_proj≥0.90). It was defined for all eleven accelerating treatments and undefined for all thirteen decelerating treatments — a perfect correspondence arising directly from the model’s algebra and therefore available as an instantaneous check from the sign of Halone. HM-FOKM’s added parameter therefore carries a genuine identifiability cost, yet the model remains structurally simple and can be used with the same confidence as the classical first-order model it extends, provided monitoring programs are designed with these constraints in mind.

Authors

Institutions

Publication Details

Journal
Research Square
Published
2026-11-13
DOI
https://doi.org/10.21203/rs.3.rs-10957757/v1
Primary Topic
Microbial bioremediation and biosurfactants
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Beyond First Order Kinetics: Sensitivity, Identifiability and the Limits of An Extended Quadratic Version for Petroleum Hydrocarbon Bioremediation: Local, Global and Profile-Likelihood Analysis

Yusuf Momoh
2 links
Research Square
Microbial bioremediation and biosurfactants
preprint

Beyond First Order Kinetics: Sensitivity, Identifiability and the Limits of An Extended Quadratic Version for Petroleum Hydrocarbon Bioremediation: Local, Global and Profile-Likelihood Analysis

Yusuf Momoh
preprint en

Abstract

Abstract A companion study showed that an extended version of the first-order kinetic model, the Hindrance-Modified First-Order Kinetic Model (HM-FOKM), B(t)=1-exp⁡(-Kt+H/2 t^2 ), outperforms classical first-order kinetics on AICc in 15 of 24 treatments that exhibit genuine quadratic curvature. This study assesses the sensitivity and identifiability of this model through local and global sensitivity analysis, profile-likelihood identifiability analysis, and a new diagnostic: the time to 90 % removal (t_90). The local sensitivity functions S_Kand S_Hmaintain fixed signs throughout the domain (S_K≥0, S_H≤0), a simpler structure than that of models whose sensitivities change sign. Because Kand Hare strongly correlated (mean ∣ρ∣=0.924), independent-box Sobol sampling is invalid. A correlation-ratio ensemble reveals that K dominates early (S_1 (K)=0.98 at 1 week), while Hovertakes Kat approximately 9.8 weeks Profile-likelihood analysis confirms that Kis better constrained than H. The 95 % confidence interval for Hincludes zero in 10 of 24 treatments (41.7 %), so classical first-order kinetics cannot be excluded even when HM-FOKM is preferred by AICc. The correlation ∣ρ(Kⓜ,H)∣ranges from 0.890 to 0.982, consistent with a general collinearity pattern in two-parameter saturating kinetic models. The t_90diagnostic exists only when K^2≥2H⋅ln⁡10(equivalently B_proj≥0.90). It was defined for all eleven accelerating treatments and undefined for all thirteen decelerating treatments — a perfect correspondence arising directly from the model’s algebra and therefore available as an instantaneous check from the sign of Halone. HM-FOKM’s added parameter therefore carries a genuine identifiability cost, yet the model remains structurally simple and can be used with the same confidence as the classical first-order model it extends, provided monitoring programs are designed with these constraints in mind.

Research Square
University of Port Harcourt (NG)
Microbial bioremediation and biosurfactants
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon