Estimation of plant litter pools and decomposition-related parameters in a mechanistic model

TY - JOUR

T1 - Estimation of plant litter pools and decomposition-related parameters in a mechanistic model

AU - Borgen, Signe Kyndig

AU - Molstad, Lars

AU - Bruun, Sander

AU - Breland, Tor Arvid

AU - Bakken, Lars Reier

AU - Blekken, Marina Azzaroli

PY - 2011

Y1 - 2011

N2 - Modeling the C and N dynamics during decomposition of plant residues depends on robust estimation of the litter pool partitioning (LPP) of residues, i. e. the fraction of their C and N allocated to rapidly and slowly decomposing pools. We searched for ways to estimate LPP by analyzing data on C and N mineralization during decomposition of 60 widely different plant residues, using a simple model with two litter pools and one microbial pool. LPP and relevant global parameters were estimated by model optimization (Levenberg-Marquardt-least-squares algorithm) in one operation. These kinetically-defined LPP values were used in regression analyses against data from stepwise chemical digestion (SCD) and near-infrared reflectance (NIR) analysis of the plant residues. LPP predicted by these regression models resulted in better performance than LPP from measured neutral-detergent-soluble (NDS) C and N when validated on independent data (n = 15 plant residues). The results demonstrated the potential improvement by simultaneous estimation of residue specific LPP and global parameters, and that kinetically-defined LPP can be equally well predicted by NIR as by total N and NDS-C. Model failure for a minority of the plant-residues could partly be removed by altering the microbial C/N ratio (global optimum 7. 1) within the range 5-15, possibly reflecting a variable dominance of bacteria or fungi.

AB - Modeling the C and N dynamics during decomposition of plant residues depends on robust estimation of the litter pool partitioning (LPP) of residues, i. e. the fraction of their C and N allocated to rapidly and slowly decomposing pools. We searched for ways to estimate LPP by analyzing data on C and N mineralization during decomposition of 60 widely different plant residues, using a simple model with two litter pools and one microbial pool. LPP and relevant global parameters were estimated by model optimization (Levenberg-Marquardt-least-squares algorithm) in one operation. These kinetically-defined LPP values were used in regression analyses against data from stepwise chemical digestion (SCD) and near-infrared reflectance (NIR) analysis of the plant residues. LPP predicted by these regression models resulted in better performance than LPP from measured neutral-detergent-soluble (NDS) C and N when validated on independent data (n = 15 plant residues). The results demonstrated the potential improvement by simultaneous estimation of residue specific LPP and global parameters, and that kinetically-defined LPP can be equally well predicted by NIR as by total N and NDS-C. Model failure for a minority of the plant-residues could partly be removed by altering the microbial C/N ratio (global optimum 7. 1) within the range 5-15, possibly reflecting a variable dominance of bacteria or fungi.

U2 - 10.1007/s11104-010-0404-4

DO - 10.1007/s11104-010-0404-4

M3 - Journal article

SN - 0032-079X

VL - 338

SP - 205

EP - 222

JO - Plant and Soil

JF - Plant and Soil

IS - 1-2

ER -