Work

Approximating Cartan subalgebra decompositions in semi simple lie algebras with applications to Yenisei taiga carbon sequestration

Year: 2025
Type: article
Abstract: Cartan subalgebra decompositions partition semisimple Lie algebras into root spaces whose dimensions and interrelations encode the algebra’s entire structure. This paper explores whether such decompositions, applied as a spectral tool, can improve predictions of carbon sequestration rates in the Yenisei River taiga one of the world’s largest boreal forest carbon sinks. A high-dimensional dataset of 23 forest inventory variables was projected onto root-space eigenvectors derived from the Lie algebra sl(6,R), yielding a reduced basis that captured 91.7% of total variance in four components. Fourier series and Bayesian prior selection complemented the algebraic decomposition by modelling seasonal periodicity and handling sparse ground-truth observations. Riemann surface theory provided a conformal mapping of the spatially irregular sampling grid onto a uniform domain. Calibrated on 12 years of eddy covariance and inventory data (2010-2021) and validated on 2022-2023 measurements, the Cartan-based model achieved RMSE of 0.37 tC/ha/yr against a mean sequestration rate of 2.14 tC/ha/yr. This outperformed principal component regression (RMSE = 0.52) and random forest (RMSE = 0.48) benchmarks. Homological analysis of the root system identified structural redundancies that allowed further dimensionality reduction without loss of predictive accuracy.
Language: English
Open Access status: diamond
OpenAlex ID: W7166315963
FWCI:
Cites: 14
Cited by:
Related to: 10