Refinement of rooting depths using satellite-based evapotranspiration seasonality for ecosystem modeling in California

Ichii, Kazuhito; Wang, Weile; Hashimoto, Hirofumi; Yang, Fan; Votava, P.; Michaelis, Andrew; Nemani, Ramakrishna R.

doi:10.1016/j.agrformet.2009.06.019

Cited by 60 publications

(45 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Support Vector Machine regression is a machine learning technique that transforms nonlinear regressions into linear regressions by mapping the original low-dimensional input space to a higher-dimensional feature space using kernel functions (e.g., Cristianini and Shawe-Taylor, 2000). The method was assessed for more than 20 Ameriflux sites over the continental United States to estimate spatial distributions both in evapotranspiration (Yang et al, 2006;Ichii et al, 2009) and in gross primary productivity (Yang et al, 2007). The model calculates GPP only as carbon cycle component.…”

Section: A11 Support Vector Machine-based Regression (Yang Et Al 2mentioning

confidence: 99%

“…The model calculates GPP only as carbon cycle component. The model output has also been used for inverse estimation of key biosphere model parameters, such as maximum light use efficiency (Yang et al, 2007) and the rooting depth (Ichii et al, 2009), and for the analysis of climate and terrestrial carbon cycles in Asia . As an original model, we used the model tuned for the AmeriFlux observation sites, which were substantially similar to those studied by Yang et al (2007).…”

Section: A11 Support Vector Machine-based Regression (Yang Et Al 2mentioning

confidence: 99%

See 1 more Smart Citation

Multi-model analysis of terrestrial carbon cycles in Japan: limitations and implications of model calibration using eddy flux observations

et al. 2010

View full text Add to dashboard Cite

Abstract. Terrestrial biosphere models show large differences when simulating carbon and water cycles, and reducing these differences is a priority for developing more accurate estimates of the condition of terrestrial ecosystems and future climate change. To reduce uncertainties and improve the understanding of their carbon budgets, we investigated the utility of the eddy flux datasets to improve model simulations and reduce variabilities among multi-model outputs of terrestrial biosphere models in Japan. Using 9 terrestrial biosphere models (Support Vector Machine -based regressions, TOPS, CASA, VISIT, Biome-BGC, DAYCENT, SEIB, LPJ, and TRIFFID), we conducted two simulations: (1) point simulations at four eddy flux sites in Japan and (2) spatial simulations for Japan with a default model (based on original settings) and a modified model (based on model parameter tuning using eddy flux data). Generally, models using default model settings showed large deviations in model outputs from observation with large model-by-model variability. However, after we calibrated the model parameters using eddy flux data (GPP, RE and NEP), most models successfully simulated seasonal variations in the carbon cycle, with less variability among models. We also found that interannualCorrespondence to: K. Ichii (kazuhito.ichii@gmail.com) variations in the carbon cycle are mostly consistent among models and observations. Spatial analysis also showed a large reduction in the variability among model outputs. This study demonstrated that careful validation and calibration of models with available eddy flux data reduced model-bymodel differences. Yet, site history, analysis of model structure changes, and more objective procedure of model calibration should be included in the further analysis.

show abstract

Section: A11 Support Vector Machine-based Regression (Yang Et Al 2mentioning

confidence: 99%

Section: A11 Support Vector Machine-based Regression (Yang Et Al 2mentioning

confidence: 99%

Multi-model analysis of terrestrial carbon cycles in Japan: limitations and implications of model calibration using eddy flux observations

et al. 2010

View full text Add to dashboard Cite

show abstract

“…Within California, average annual temperature is predicted to increase by up to 5°C by 2100, resulting in projected increases in water deficit of 30% or more in many areas (absolute increases of up to 100 mm in annual deficit), greatly exceeding the estimated historical changes in this study (8,50). However, estimates of CWD and our understanding of how plants experience CWD are limited by our understanding of interactions between water storage and the ability of plants to access deeper reservoirs of water (30,51). Trees in montane areas may avoid deficit by tapping water available below the depth of most current soil models (30), and some areas of California are predicted to exhibit declines in water deficit over the next century due to increases in precipitation in areas with high soil water storage potential (e.g., ref.…”

Section: Discussionmentioning

confidence: 99%

“…We regressed changes in tree numbers against change in CWD within each tree size class. To address the complexity of soils and estimates of CWD in the montane Sierra Nevada, where forest trees may access water reservoirs deeper than those included in our CWD model (30,51), we examined the effect of excluding the Sierra Nevada from our analyses. Results were robust to exclusion of the Sierra Nevada (SI Appendix, Table S5 and Fig.…”

Section: Methodsmentioning

confidence: 99%

California's Historic Legacy For Landscape Change, the Wieslander Vegetation Type Maps

Thorne

2016

Madroño

View full text Add to dashboard Cite

We document changes in forest structure between historical (1930s) and contemporary (2000s) surveys of California vegetation through comparisons of tree abundance and size across the state and within several ecoregions. Across California, tree density in forested regions increased by 30% between the two time periods, whereas forest biomass in the same regions declined, as indicated by a 19% reduction in basal area. These changes reflect a demographic shift in forest structure: larger trees (>61 cm diameter at breast height) have declined, whereas smaller trees (<30 cm) have increased. Large tree declines were found in all surveyed regions of California, whereas small tree increases were found in every region except the south and central coast. Large tree declines were more severe in areas experiencing greater increases in climatic water deficit since the 1930s, based on a hydrologic model of water balance for historical climates through the 20th century. Forest composition in California in the last century has also shifted toward increased dominance by oaks relative to pines, a pattern consistent with warming and increased water stress, and also with paleohistoric shifts in vegetation in California over the last 150,000 y.global change | forest | historical ecology | climatic water deficit U nderstanding the patterns and causes of recent changes in vegetation structure is essential to predicting ongoing and future vegetation responses to global climate change. Recent changes in forest structure and tree mortality attributed to increases in temperature and drought have been documented for large areas across the globe (1-5). Declines in the abundance of large trees have attracted particular attention, as large trees contribute disproportionately to forest structure and function, carbon stocks, and the cultural values of forests (6-10). Although land-use change and harvesting of large trees contribute to their decline, studies have found that large trees can suffer disproportionate mortality in response to drought in both temperate and tropical systems, and that declines have occurred in protected areas not subject to logging (5,8,11,12). The mechanisms responsible for large tree vulnerability to water deficit are poorly understood but rest largely on interacting effects of increased vulnerability to cavitation, carbon starvation during drought, and vulnerability to natural enemies (13,14). Here we present, to our knowledge, the first analysis of changes in large tree density across California during the 20th century in relation to changes in water deficit during the same period, and examine changes in species composition that can be linked to historical changes over longer time periods.In the forests of California, comparisons of historical and contemporary forests in selected areas of the Sierra Nevada and Transverse mountains suggest that forests have in-filled with small trees and exhibited declines in larger trees since the early 1900s, patterns attributed to fire suppression, forest exploitation, and changes ...

show abstract

“…The use of these satellite-derived products to verify LSM simulations or to optimise key LSM parameters has been assessed by several authors (e.g. Becker-Reshef et al, 2010;Crow et al, 2012;Ferrant et al, 2014;Ford et al, 2014;Ghilain et al, 2012;Ichii et al, 2009;Kowalik et al, 2014;Szczypta et al, 2012Szczypta et al, , 2014. Data assimilation is a field of active research.…”

Section: Introductionmentioning

confidence: 99%

Parameter optimisation for a better representation of drought by LSMs: inverse modelling vs. sequential data assimilation

Dewaele¹,

Munier²,

Albergel³

et al. 2017

Hydrol. Earth Syst. Sci.

View full text Add to dashboard Cite

Abstract. Soil maximum available water content (Max-AWC) is a key parameter in land surface models (LSMs). However, being difficult to measure, this parameter is usually uncertain. This study assesses the feasibility of using a 15-year (1999-2013) time series of satellite-derived lowresolution observations of leaf area index (LAI) to estimate MaxAWC for rainfed croplands over France. LAI interannual variability is simulated using the CO 2 -responsive version of the Interactions between Soil, Biosphere and Atmosphere (ISBA) LSM for various values of MaxAWC. Optimal value is then selected by using (1) a simple inverse modelling technique, comparing simulated and observed LAI and (2) a more complex method consisting in integrating observed LAI in ISBA through a land data assimilation system (LDAS) and minimising LAI analysis increments. The evaluation of the MaxAWC estimates from both methods is done using simulated annual maximum above-ground biomass (B ag ) and straw cereal grain yield (GY) values from the Agreste French agricultural statistics portal, for 45 administrative units presenting a high proportion of straw cereals. Significant correlations (p value < 0.01) between B ag and GY are found for up to 36 and 53 % of the administrative units for the inverse modelling and LDAS tuning methods, respectively. It is found that the LDAS tuning experiment gives more realistic values of MaxAWC and maximum B ag than the inverse modelling experiment. Using undisaggregated LAI observations leads to an underestimation of MaxAWC and maximum B ag in both experiments. Median annual maximum values of disaggregated LAI observations are found to correlate very well with MaxAWC.

show abstract

Refinement of rooting depths using satellite-based evapotranspiration seasonality for ecosystem modeling in California

Cited by 60 publications

References 53 publications

Multi-model analysis of terrestrial carbon cycles in Japan: limitations and implications of model calibration using eddy flux observations

Multi-model analysis of terrestrial carbon cycles in Japan: limitations and implications of model calibration using eddy flux observations

California's Historic Legacy For Landscape Change, the Wieslander Vegetation Type Maps

Parameter optimisation for a better representation of drought by LSMs: inverse modelling vs. sequential data assimilation

Contact Info

Product

Resources

About