Aki, Tsuruta: Spatial and temporal distribution of European CH₄ emissions from process-based models and CTE-CH₄ atmospheric inverse model
0 likes46 views
1) Spatial and temporal distributions of European CH4 emissions from bottom-up process models and the CTE-CH4 atmospheric inversion model show significant variations. 2) Inversions using two different wetland prior emissions (LPX-Bern and JSBACH-HIMMELI) reduced differences in northern Europe but differences remained in other regions due to lack of atmospheric observations. 3) Both prior models and posterior estimates show small winter wetland emissions and a summer maximum in northern Europe, but magnitudes and timing of peaks differed between the models.
![3
Background
European emissions: seasonal cycle
●
Large uncertainty in emissions from
wetlands
– Seasonal cycle amplitude (SCA) vary
significantly by different inversions and
process-based models.
– Monthly median from TD shows very
small SCA, while 95th percentile (upper
limit) show amplitude of approx. 0.7 Tg
CH4 month-1
– BU SCA tends to be higher than that of
TD
– Some BU models show high winter-
spring emissions, close to summer level
Average monthly European* fluxes during
2008-2017
Top-down (TD) Bottom-up (BU)
●
European domain: [35°N-73°N, 13°W-38°E]
●
Prognostic: models used their own internal approach
to estimate wetland area
●
Diagnostic: wetland surface areas from Wetland Area
Dynamics for Methane Modeling (WAD2M)
Solid line: median of model ensemble, Dotted lines: individual model
Shaded areas: between 5th and 95th percentiles](https://image.slidesharecdn.com/tsurutaakisession18-201117084536/85/Aki-Tsuruta-Spatial-and-temporal-distribution-of-European-CH-emissions-from-process-based-models-and-CTE-CH-atmospheric-inverse-model-3-320.jpg)
Aki, Tsuruta: Spatial and temporal distribution of European CH₄ emissions from process-based models and CTE-CH₄ atmospheric inverse model
- 1. Spatial and temporal distribution of European CH4 emissions from process- based models and CTE-CH4 atmospheric inverse model Tsuruta Aki1 , Leif Backman1 , Tiina Markkanen1 , Maarit Raivonen2 , Antti Leppänen2 , Sebastian Lienert3 , Fortunat Joos3 , Jurek Müller3 , Hugo Denier van der Gon4 , Janssens-Maenhout Greet5 , European and global atm. CH4 station PIs, Tuula Aalto1 17/09/2020 ICOS Science Conference 2020, Online
- 2. 2 Background European emissions ● Largest contribution from agriculture and waste sectors ● Second most from fossil fuel production and use ● Emissions from wetlands are the largest natural source Source categ EU Map
- 3. 3 Background European emissions: seasonal cycle ● Large uncertainty in emissions from wetlands – Seasonal cycle amplitude (SCA) vary significantly by different inversions and process-based models. – Monthly median from TD shows very small SCA, while 95th percentile (upper limit) show amplitude of approx. 0.7 Tg CH4 month-1 – BU SCA tends to be higher than that of TD – Some BU models show high winter- spring emissions, close to summer level Average monthly European* fluxes during 2008-2017 Top-down (TD) Bottom-up (BU) ● European domain: [35°N-73°N, 13°W-38°E] ● Prognostic: models used their own internal approach to estimate wetland area ● Diagnostic: wetland surface areas from Wetland Area Dynamics for Methane Modeling (WAD2M) Solid line: median of model ensemble, Dotted lines: individual model Shaded areas: between 5th and 95th percentiles
- 4. 4 Background European emissions: spatial distribution ● High anthropogenic emissions in cities, agricultural areas → high in central Europe – TD estimates do not vary so significantly between models ● Biospheric emissions are high in northern and north-east Europe – Locations of hot spots vary much between TD, BU-Prognostic and BU- Diagnostic – Rage in estimates is significantly higher than that of anthropogenic emissions (Max. - Min.) / MeanMean TD Anthropogenic TD Biospheric BU Prognostic BU Diagnostic TD Anthropogenic BU Prognostic BU Diagnostic TD Biospheric Mean and rage of CH4 emission estimates over Europe, 2005-2017 average *Mean of model ensembles, is calculated from 2005-2017 monthly data. *Min. and Max. is minimum and maximum of model ensembles.
- 5. 5 ● Optimize European CH4 using CarbonTracker Europe-CH4 atmospheric inverse model – Grid-based optimization over Europe: 1° x 1°, 3° x 2°, 6° x 4° – Spatial correlation: 100-500 km ● Use two distinct sets of wetland priors emissions ● LPX-Bern v1.4 (global, orig. resolution 0.5° x 0.5° x monthly) (Lienert and Joos, 2018) – Inundated wetlands, wet soil, soil sinks, peatlands – Wetland/vegetation distributions calculated with DYPTOP model ● JSBACH-HIMMELI (Europe only, orig. resolution 0.1° x 0.1° x daily) (Raivonen et al., 2017) – Mineral soils (can be sinks or sources), peatlands – Wetland/vegetation distributions based on EU Corine Methods 1x1, 100 km 3x2, 200 km 6x4, 500 km
- 7. 7 ● Atmospheric CH4 as constraints: mainly NOAA + ICOS observations over Europe ● Good spatial coverage of continuous stations over central and northern Europe, especially for late years ● Continuous hourly data are pre- processed before inversion: – Filtered by taking only “good quality” observations – Afternoon 12-16 LT averages – Night time 0-4 LT averages for mountain sites Atmospheric CH4 observations Locations of atmospheric CH4 observational sites, data available from 2000-2018
- 8. 8 ● Prior differences – Largest in the northern peatland area, and hot-spot in Scotland and east Hungary ● Posterior – Northern peatland area: increase from LPX-Bern v1.4, decrease from JSBACH-HIMMELI → differences are smaller than the prior – Central and Southern Europe: JSBACH-HIMMELI negative fluxes are smaller in posterior, but an increase in LPX-Bern v1.4 emissions → differences still remain/increase from the prior Results Biospheric (wetlands as net total) flux estimates and their differences, 2005 mean Posterior Prior
- 9. 9 ● Anthropogenic emissions – Same priors are used (EDGAR v5.0) – Higher emissions in the inversion using JSBACH-HIMMELI at central Europe, i.e. compensating effect from the biospheric emissions. Results Anthropogenic emission estimates and their differences, 2005 mean Posterior Prior EDGAR v5.0 EDGAR v5.0
- 10. 10 ● Anthropogenic emissions – Same priors are used (EDGAR v5.0) – Higher emissions in the inversion using JSBACH-HIMMELI at central Europe, i.e. compensating effect from the biospheric emissions. ● Total emissions – Differences are reduced for northern Europe, and slightly in central Europe – Differences of hot-spots (Scotland and east Hungary) remains – Differences in the north eastern Europe remain Results Total emission estimates and their differences, 2005 mean Posterior Prior
- 11. 11 Results Biospheric emissions seasonal cycle ● Europe, 60°N> – Both priors and posteriors show small winter emissions and clear emission maximum in summer – Seasonal max. in LPX-Bern v1.4 is much later than of JSBACH-HIMMELI → after inversion, both has summer max. in August. – Emission magnitudes are close in posterior, but still differ approx. twice. ● Whole European domain – Very different shape of seasonal cycle in the prior and posterior Monthly total biospheric fluxes, 2005 Whole European domain Europe, 60°N> Solid lines: posterior, dashed lines: prior
- 12. 12 Conclusion ● Seasonal cycle and spatial distribution of European emissions vary significantly by different bottom-up and top-down estimates – Wetlands are the main source of uncertainty ● CTE-CH4 could constrain the wetlands emissions for northern Europe to some extent, but had difficulties in east Europe due to luck of atmospheric observation. ● Soil sink is small as magnitude, hence is difficult for inversion to turn signs → leads to compensating changes in other emission sources ● Way forward – Not all process-based models estimate wetland extent & distribution themselves, or take their temporal changes into account. → Consider such as one uncertainty in inversion or a parameter to be constrained – Isotope observations (δ13 C-CH4) can be considered as additional source of constraint, but they are still limited in numbers and frequencies, or yet synthesized between laboratories.
- 13. 13 Global atm. CH4 sites
- 14. 14 Mean flux scaling factors, biospheric Mean flux scaling factors, anthropogenic