A new scheme for initializing process-based ecosystem models by scaling soil carbon pools

S. Hashimoto, M. Wattenbach, P. Smith

Research output: Contribution to journalArticle

19 Citations (Scopus)

Abstract

Process-based ecosystem models are useful tools, not only for understanding the forest carbon cycle, but also for predicting future change. In order to apply a model to simulate a specific time period, model initialization is required. In this study, we propose a new scheme of initialization for forest ecosystem models, which we term a "slow-relaxation scheme", that entails scaling of the soil carbon and nitrogen pools slowly during the spin-up period. The proposed slow-relation scheme was tested with the CENTURY version 4 ecosystem model. Three different combinations of scaled soil pools were also tested, and compared to the results from a fast-relaxation regime. The fast-relaxation of soil pools produced unstable, transient model behaviour whereas slow-relaxation overcame this instability. This approach holds promise for initializing ecosystem models, and for starting simulations with more realistic initial conditions. (C) 2011 Elsevier B.V. All rights reserved.

Original languageEnglish
Pages (from-to)3598-3602
Number of pages5
JournalEcological Modelling
Volume222
Issue number19
Early online date28 Sep 2011
DOIs
Publication statusPublished - 10 Oct 2011

Keywords

  • Ecosystem model
  • Initialization
  • Spin-up
  • Soil carbon
  • Forest
  • Nitrogen
  • ROTHC
  • Cycle

Cite this

A new scheme for initializing process-based ecosystem models by scaling soil carbon pools. / Hashimoto, S.; Wattenbach, M.; Smith, P.

In: Ecological Modelling, Vol. 222, No. 19, 10.10.2011, p. 3598-3602.

Research output: Contribution to journalArticle

@article{721cc445ddf54fe78eb791bb3f72ad34,
title = "A new scheme for initializing process-based ecosystem models by scaling soil carbon pools",
abstract = "Process-based ecosystem models are useful tools, not only for understanding the forest carbon cycle, but also for predicting future change. In order to apply a model to simulate a specific time period, model initialization is required. In this study, we propose a new scheme of initialization for forest ecosystem models, which we term a {"}slow-relaxation scheme{"}, that entails scaling of the soil carbon and nitrogen pools slowly during the spin-up period. The proposed slow-relation scheme was tested with the CENTURY version 4 ecosystem model. Three different combinations of scaled soil pools were also tested, and compared to the results from a fast-relaxation regime. The fast-relaxation of soil pools produced unstable, transient model behaviour whereas slow-relaxation overcame this instability. This approach holds promise for initializing ecosystem models, and for starting simulations with more realistic initial conditions. (C) 2011 Elsevier B.V. All rights reserved.",
keywords = "Ecosystem model, Initialization, Spin-up, Soil carbon, Forest, Nitrogen, ROTHC, Cycle",
author = "S. Hashimoto and M. Wattenbach and P. Smith",
year = "2011",
month = "10",
day = "10",
doi = "10.1016/j.ecolmodel.2011.08.011",
language = "English",
volume = "222",
pages = "3598--3602",
journal = "Ecological Modelling",
issn = "0304-3800",
publisher = "Elsevier Science B. V.",
number = "19",

}

TY - JOUR

T1 - A new scheme for initializing process-based ecosystem models by scaling soil carbon pools

AU - Hashimoto, S.

AU - Wattenbach, M.

AU - Smith, P.

PY - 2011/10/10

Y1 - 2011/10/10

N2 - Process-based ecosystem models are useful tools, not only for understanding the forest carbon cycle, but also for predicting future change. In order to apply a model to simulate a specific time period, model initialization is required. In this study, we propose a new scheme of initialization for forest ecosystem models, which we term a "slow-relaxation scheme", that entails scaling of the soil carbon and nitrogen pools slowly during the spin-up period. The proposed slow-relation scheme was tested with the CENTURY version 4 ecosystem model. Three different combinations of scaled soil pools were also tested, and compared to the results from a fast-relaxation regime. The fast-relaxation of soil pools produced unstable, transient model behaviour whereas slow-relaxation overcame this instability. This approach holds promise for initializing ecosystem models, and for starting simulations with more realistic initial conditions. (C) 2011 Elsevier B.V. All rights reserved.

AB - Process-based ecosystem models are useful tools, not only for understanding the forest carbon cycle, but also for predicting future change. In order to apply a model to simulate a specific time period, model initialization is required. In this study, we propose a new scheme of initialization for forest ecosystem models, which we term a "slow-relaxation scheme", that entails scaling of the soil carbon and nitrogen pools slowly during the spin-up period. The proposed slow-relation scheme was tested with the CENTURY version 4 ecosystem model. Three different combinations of scaled soil pools were also tested, and compared to the results from a fast-relaxation regime. The fast-relaxation of soil pools produced unstable, transient model behaviour whereas slow-relaxation overcame this instability. This approach holds promise for initializing ecosystem models, and for starting simulations with more realistic initial conditions. (C) 2011 Elsevier B.V. All rights reserved.

KW - Ecosystem model

KW - Initialization

KW - Spin-up

KW - Soil carbon

KW - Forest

KW - Nitrogen

KW - ROTHC

KW - Cycle

U2 - 10.1016/j.ecolmodel.2011.08.011

DO - 10.1016/j.ecolmodel.2011.08.011

M3 - Article

VL - 222

SP - 3598

EP - 3602

JO - Ecological Modelling

JF - Ecological Modelling

SN - 0304-3800

IS - 19

ER -