Confronting terrestrial biosphere models with forest inventory data

Jeremy W. Lichstein, Ni Zhang Golaz, Sergey Malyshev, Elena Shevliakova, Tao Zhang, Justin Sheffield, Richard A. Birdsey, Jorge Louis Sarmiento, Stephen Wilson Pacala

Research output: Contribution to journalArticle

16 Citations (Scopus)

Abstract

Efforts to test and improve terrestrial biosphere models (TBMs) using a variety of data sources have become increasingly common. Yet, geographically extensive forest inventories have been under-exploited in previous model-data fusion efforts. Inventory observations of forest growth, mortality, and biomass integrate processes across a range of timescales, including slow timescale processes such as species turnover, that are likely to have important effects on ecosystem responses to environmental variation. However, the large number (thousands) of inventory plots precludes detailed measurements at each location, so that uncertainty in climate, soil properties, and other environmental drivers may be large. Errors in driver variables, if ignored, introduce bias into model-data fusion. We estimated errors in climate and soil drivers at U.S. Forest Inventory and Analysis (FIA) plots, and we explored the effects of these errors on model-data fusion with the Geophysical Fluid Dynamics Laboratory LM3V dynamic global vegetation model. When driver errors were ignored or assumed small at FIA plots, responses of biomass production in LM3V to precipitation and soil available water capacity appeared steeper than the corresponding responses estimated from FIA data. These differences became nonsignificant if driver errors at FIA plots were assumed to be large. Ignoring driver errors when optimizing LM3V parameter values yielded estimates for fine-root allocation that were larger than biometric estimates, which is consistent with the expected direction of bias. To explore whether complications posed by driver errors could be circumvented by relying on intensive study sites where driver errors are small, we performed a power analysis. To accurately quantify the response of biomass production to spatial variation in mean annual precipitation within the eastern United States would require at least 40 intensive study sites, which is larger than the number of sites typically available for individual biomes in existing plot networks. Driver errors may be accommodated by several existing model-data fusion approaches, including hierarchical Bayesian methods and ensemble filtering methods; however, these methods are computationally expensive. We propose a new approach, in which the TBM functional response is fit directly to the driver-error-corrected functional response estimated from data, rather than to the raw observations.

Original languageEnglish (US)
Pages (from-to)699-715
Number of pages17
JournalEcological Applications
Volume24
Issue number4
DOIs
StatePublished - Jun 2014

Fingerprint

forest inventory
biosphere
functional response
biomass
timescale
available water capacity
ecosystem response
biometry
response analysis
fluid dynamics
climate
fine root
biome
soil property
turnover
spatial variation
soil
mortality
vegetation
method

All Science Journal Classification (ASJC) codes

  • Ecology

Keywords

  • Carbon cycle model
  • Data assimilation
  • Errors in explanatory variables
  • Global ecosystem model
  • Land surface model
  • Measurement error models

Cite this

Lichstein, J. W., Golaz, N. Z., Malyshev, S., Shevliakova, E., Zhang, T., Sheffield, J., ... Pacala, S. W. (2014). Confronting terrestrial biosphere models with forest inventory data. Ecological Applications, 24(4), 699-715. https://doi.org/10.1890/13-0600.1
Lichstein, Jeremy W. ; Golaz, Ni Zhang ; Malyshev, Sergey ; Shevliakova, Elena ; Zhang, Tao ; Sheffield, Justin ; Birdsey, Richard A. ; Sarmiento, Jorge Louis ; Pacala, Stephen Wilson. / Confronting terrestrial biosphere models with forest inventory data. In: Ecological Applications. 2014 ; Vol. 24, No. 4. pp. 699-715.
@article{4a1e0c2adbc048afab0162b87c259cf7,
title = "Confronting terrestrial biosphere models with forest inventory data",
abstract = "Efforts to test and improve terrestrial biosphere models (TBMs) using a variety of data sources have become increasingly common. Yet, geographically extensive forest inventories have been under-exploited in previous model-data fusion efforts. Inventory observations of forest growth, mortality, and biomass integrate processes across a range of timescales, including slow timescale processes such as species turnover, that are likely to have important effects on ecosystem responses to environmental variation. However, the large number (thousands) of inventory plots precludes detailed measurements at each location, so that uncertainty in climate, soil properties, and other environmental drivers may be large. Errors in driver variables, if ignored, introduce bias into model-data fusion. We estimated errors in climate and soil drivers at U.S. Forest Inventory and Analysis (FIA) plots, and we explored the effects of these errors on model-data fusion with the Geophysical Fluid Dynamics Laboratory LM3V dynamic global vegetation model. When driver errors were ignored or assumed small at FIA plots, responses of biomass production in LM3V to precipitation and soil available water capacity appeared steeper than the corresponding responses estimated from FIA data. These differences became nonsignificant if driver errors at FIA plots were assumed to be large. Ignoring driver errors when optimizing LM3V parameter values yielded estimates for fine-root allocation that were larger than biometric estimates, which is consistent with the expected direction of bias. To explore whether complications posed by driver errors could be circumvented by relying on intensive study sites where driver errors are small, we performed a power analysis. To accurately quantify the response of biomass production to spatial variation in mean annual precipitation within the eastern United States would require at least 40 intensive study sites, which is larger than the number of sites typically available for individual biomes in existing plot networks. Driver errors may be accommodated by several existing model-data fusion approaches, including hierarchical Bayesian methods and ensemble filtering methods; however, these methods are computationally expensive. We propose a new approach, in which the TBM functional response is fit directly to the driver-error-corrected functional response estimated from data, rather than to the raw observations.",
keywords = "Carbon cycle model, Data assimilation, Errors in explanatory variables, Global ecosystem model, Land surface model, Measurement error models",
author = "Lichstein, {Jeremy W.} and Golaz, {Ni Zhang} and Sergey Malyshev and Elena Shevliakova and Tao Zhang and Justin Sheffield and Birdsey, {Richard A.} and Sarmiento, {Jorge Louis} and Pacala, {Stephen Wilson}",
year = "2014",
month = "6",
doi = "10.1890/13-0600.1",
language = "English (US)",
volume = "24",
pages = "699--715",
journal = "Ecological Applications",
issn = "1051-0761",
publisher = "Ecological Society of America",
number = "4",

}

Lichstein, JW, Golaz, NZ, Malyshev, S, Shevliakova, E, Zhang, T, Sheffield, J, Birdsey, RA, Sarmiento, JL & Pacala, SW 2014, 'Confronting terrestrial biosphere models with forest inventory data', Ecological Applications, vol. 24, no. 4, pp. 699-715. https://doi.org/10.1890/13-0600.1

Confronting terrestrial biosphere models with forest inventory data. / Lichstein, Jeremy W.; Golaz, Ni Zhang; Malyshev, Sergey; Shevliakova, Elena; Zhang, Tao; Sheffield, Justin; Birdsey, Richard A.; Sarmiento, Jorge Louis; Pacala, Stephen Wilson.

In: Ecological Applications, Vol. 24, No. 4, 06.2014, p. 699-715.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Confronting terrestrial biosphere models with forest inventory data

AU - Lichstein, Jeremy W.

AU - Golaz, Ni Zhang

AU - Malyshev, Sergey

AU - Shevliakova, Elena

AU - Zhang, Tao

AU - Sheffield, Justin

AU - Birdsey, Richard A.

AU - Sarmiento, Jorge Louis

AU - Pacala, Stephen Wilson

PY - 2014/6

Y1 - 2014/6

N2 - Efforts to test and improve terrestrial biosphere models (TBMs) using a variety of data sources have become increasingly common. Yet, geographically extensive forest inventories have been under-exploited in previous model-data fusion efforts. Inventory observations of forest growth, mortality, and biomass integrate processes across a range of timescales, including slow timescale processes such as species turnover, that are likely to have important effects on ecosystem responses to environmental variation. However, the large number (thousands) of inventory plots precludes detailed measurements at each location, so that uncertainty in climate, soil properties, and other environmental drivers may be large. Errors in driver variables, if ignored, introduce bias into model-data fusion. We estimated errors in climate and soil drivers at U.S. Forest Inventory and Analysis (FIA) plots, and we explored the effects of these errors on model-data fusion with the Geophysical Fluid Dynamics Laboratory LM3V dynamic global vegetation model. When driver errors were ignored or assumed small at FIA plots, responses of biomass production in LM3V to precipitation and soil available water capacity appeared steeper than the corresponding responses estimated from FIA data. These differences became nonsignificant if driver errors at FIA plots were assumed to be large. Ignoring driver errors when optimizing LM3V parameter values yielded estimates for fine-root allocation that were larger than biometric estimates, which is consistent with the expected direction of bias. To explore whether complications posed by driver errors could be circumvented by relying on intensive study sites where driver errors are small, we performed a power analysis. To accurately quantify the response of biomass production to spatial variation in mean annual precipitation within the eastern United States would require at least 40 intensive study sites, which is larger than the number of sites typically available for individual biomes in existing plot networks. Driver errors may be accommodated by several existing model-data fusion approaches, including hierarchical Bayesian methods and ensemble filtering methods; however, these methods are computationally expensive. We propose a new approach, in which the TBM functional response is fit directly to the driver-error-corrected functional response estimated from data, rather than to the raw observations.

AB - Efforts to test and improve terrestrial biosphere models (TBMs) using a variety of data sources have become increasingly common. Yet, geographically extensive forest inventories have been under-exploited in previous model-data fusion efforts. Inventory observations of forest growth, mortality, and biomass integrate processes across a range of timescales, including slow timescale processes such as species turnover, that are likely to have important effects on ecosystem responses to environmental variation. However, the large number (thousands) of inventory plots precludes detailed measurements at each location, so that uncertainty in climate, soil properties, and other environmental drivers may be large. Errors in driver variables, if ignored, introduce bias into model-data fusion. We estimated errors in climate and soil drivers at U.S. Forest Inventory and Analysis (FIA) plots, and we explored the effects of these errors on model-data fusion with the Geophysical Fluid Dynamics Laboratory LM3V dynamic global vegetation model. When driver errors were ignored or assumed small at FIA plots, responses of biomass production in LM3V to precipitation and soil available water capacity appeared steeper than the corresponding responses estimated from FIA data. These differences became nonsignificant if driver errors at FIA plots were assumed to be large. Ignoring driver errors when optimizing LM3V parameter values yielded estimates for fine-root allocation that were larger than biometric estimates, which is consistent with the expected direction of bias. To explore whether complications posed by driver errors could be circumvented by relying on intensive study sites where driver errors are small, we performed a power analysis. To accurately quantify the response of biomass production to spatial variation in mean annual precipitation within the eastern United States would require at least 40 intensive study sites, which is larger than the number of sites typically available for individual biomes in existing plot networks. Driver errors may be accommodated by several existing model-data fusion approaches, including hierarchical Bayesian methods and ensemble filtering methods; however, these methods are computationally expensive. We propose a new approach, in which the TBM functional response is fit directly to the driver-error-corrected functional response estimated from data, rather than to the raw observations.

KW - Carbon cycle model

KW - Data assimilation

KW - Errors in explanatory variables

KW - Global ecosystem model

KW - Land surface model

KW - Measurement error models

UR - http://www.scopus.com/inward/record.url?scp=84896044272&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84896044272&partnerID=8YFLogxK

U2 - 10.1890/13-0600.1

DO - 10.1890/13-0600.1

M3 - Article

C2 - 24988769

AN - SCOPUS:84896044272

VL - 24

SP - 699

EP - 715

JO - Ecological Applications

JF - Ecological Applications

SN - 1051-0761

IS - 4

ER -

Lichstein JW, Golaz NZ, Malyshev S, Shevliakova E, Zhang T, Sheffield J et al. Confronting terrestrial biosphere models with forest inventory data. Ecological Applications. 2014 Jun;24(4):699-715. https://doi.org/10.1890/13-0600.1