- Email: [email protected]
Assessing the resources and mitigation potential of European forests
Available online at www.sciencedirect.com
ScienceDirect ScienceDirect Energy Procedia 00 (2017) 000–000
Available online at www.sciencedirect.com Availableonline onlineatatwww.sciencedirect.com www.sciencedirect.com Available Energy Procedia 00 (2017) 000–000
ScienceDirect ScienceDirect ScienceDirect Energy Procedia 00 (2017) 000–000 Energy (2017) 000–000 372–378 EnergyProcedia Procedia125 00 (2017)
www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia
www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia
European Geosciences Union General Assembly 2017, EGU Division Energy, Resources & Environment, ERE EGU European Geosciences Union General Assembly 2017, Division Energy, Resources & Environment, ERE
Assessing theEuropean resources and mitigation potential of European forests Geosciences Union General Assembly 2017, EGU Theresources 15th International on District Heating Cooling forests Assessing the andSymposium mitigation potential ofand European Division Energy, Resources a,* a & Environment, b ERE c Hubert Hasenauer , Mathias Neumann , Adam Moreno , and Steve Running a,* a c Hubert Hasenauer ,feasibility Mathias Neumann , Adam , and82,Steve Running Assessing the of using theMoreno heat bdemand-outdoor University of Natural Resources and Life Science, Institute of Silviculture, Peter-Jordan-Str. 1180 Wien, Austria
Assessing the NASA resources and mitigation potential of European forests Ames Research Center, Mail Stop 204-14, Moffett Field, CA 94035-0001, USA University of Natural Resources and Life Science, Institute of Silviculture, Peter-Jordan-Str. 82, 1180 Wien, Austria temperature function for aof Forestry long-term district heat demand Numerical Terradynamic Simulation Group, College & Conservation, The University of Montana, Missoula, MTforecast 59812, USA a,* Research Center, Mail Stop 204-14, a b c NASA Ames Moffett Field, CA 94035-0001, USA a
b
c
a
b
c
Hubert Hasenauer , Mathias Neumann , Adam Moreno , and Steve Running a a b c I. University Andrićofa,b,c *, A. Pinaand , P. , J. Fournier B. Lacarrière , O. LeAustria Correc Natural Resources LifeFerrão Science, Institute of Silviculture,.,Peter-Jordan-Str. 82, 1180 Wien,
Numerical Terradynamic Simulation Group, College of Forestry & Conservation, The University of Montana, Missoula, MT 59812, USA a
b NASA Ames Research Center, Mail Stop 204-14, Moffett Field, CA 94035-0001, USA Abstract a IN+ Center for Innovation, Technology and Policy Research - Instituto Superior Técnico, Av. Rovisco Pais 1, 1049-001 Lisbon, Portugal c
Numerical Terradynamic Simulation Group, College of Forestry & Conservation, The University of Montana, Missoula, MT 59812, USA b
Veolia Recherche & Innovation, 291 Avenue Dreyfous Daniel, 78520 Limay, France Abstract c National and international carbon reporting systems require information on forest carbon This information can be derived Département Systèmes Énergétiques et Environnement - IMT Atlantique, 4 rue Alfredstocks. Kastler, 44300 Nantes, France from national forest inventory data and remote sensing. Here we present the conceptual challenges in assessing forest resources National and international carbon reporting systems information on estimates forest carbon stocks.from This13 information can be derived across Europe by combining MODIS satellite versusrequire terrestrial driven NPP calculated national forest inventory from national forest inventory data and remote sensing. Here we present the conceptual challenges in assessing forest resources Abstract (NFI) data covering 200.000 sampling plots. The results suggest that MODIS NPP predictions using local daily climate data and across Europe by desnsity combining MODIS satellite versus terrestrial driven estimates. NPP estimates calculated from 13leads national inventory addressing stand effects, provide realistic forest productivity Ignoring these effects to anforest overestimation Abstract (NFI) data covering 200.000 sampling plots. The results suggest that MODIS NPP predictions using local daily climate data and in the estimated carbon storage of reporting European systems forests derived satelliteon data. National and international carbon require from information forest carbon stocks. This information can be derived addressing stand desnsity effects, provide realistic forest productivity estimates. Ignoring these effects leads to an overestimation from national forest inventory data and remote sensing. Here we present the conceptual challenges in assessing forest resources in the estimated carbon storage ofMODIS; European forests derived from satellite District heating networks are commonly addressed in the literature asdata. one of the most effective solutions decreasing the Keywords: carbon; inventory; NPP; NFI, biomass; volume; MOD17; satellite; remote sensing; bioeconomy across Europe by forest combining MODIS satellite versus terrestrial driven NPP estimates calculated from 13 nationalfor forest inventory greenhouse gas emissions from the building sector. These systems require high investments which are returned through theand heat (NFI) data covering 200.000 sampling plots. The results suggest that MODIS NPP predictions using local daily climate data © 2017 The Authors. Published by Elsevier Keywords: carbon; forest inventory; MODIS; NPP;Ltd. NFI, and biomass; volume; MOD17; satellite; remote sensing; bioeconomy sales. Due to the changed climate conditions building renovation policies, heatthese demand in the future could decrease, addressing desnsity effects, provide realisticcommittee forest productivity estimates. Ignoring to an overestimation Peer-reviewstand under responsibility of the scientific of the European Geosciences Unioneffects (EGU)leads General Assembly prolonging the investment return period. in the estimated carbon storage of European forests derived from satellite data. 2017 – Division Energy, Resources and the Environment (ERE). main of this paper isbytoElsevier assess the feasibility of using the heat demand – outdoor temperature function for heat demand ©The 2017 Thescope Authors. Published Ltd. Keywords: forest inventory; MODIS; NPP; NFI, biomass; volume; MOD17; satellite; sensing; bioeconomy forecast.carbon; The district of Alvalade, in committee Lisbon (Portugal), was used as aremote case Union study.(EGU) The district consisted 2017 of 665 Peer-review under responsibility of thelocated scientific of the European Geosciences Generalis Assembly © 2017 The Authors. Published by Elsevier Ltd. buildings that vary in both construction period and typology. Three weather scenarios (low, medium, high) and three district – Division Energy, Resources and the Environment (ERE). Peer-review under responsibility of the scientific committee of the European Geosciences Union (EGU) General Assembly 2017 renovation scenarios were developed (shallow, intermediate, deep). To estimate the error, obtained heat demand values were – compared Division Energy, Resources the Environment (ERE). with results from aand dynamic heat demand model, previously developed and validated by the authors. 1. Introduction ©The 2017 The Authors. Published by Elsevier Ltd. results showed that when only weather change is considered, the margin of error could be acceptable for some applications Peer-review under responsibility of lower the scientific committee of the European (EGU)after General Assembly 2017 1. Introduction (the error in annual demand was than 20% for all weather scenariosGeosciences considered).Union However, introducing renovation 40 Energy, % of theResources European area is covered with forests managed for the provision of ecosystem services such –Today Division andland the Environment (ERE). scenarios, the error value increased up to 59.5% (depending on the weather and renovation scenarios combination considered).
asThe timber drinkingland water or ison welfare and nature Forests store large of services carbon [1, Today 40 production, %ofofslope the European area covered with forests managed for thetoprovision of amounts ecosystem such value coefficient increased average within theconservation. range of 3.8% up 8% per decade, that corresponds to 2] the and thus play a key role in the global carbon cycle including mitigating climate change effects [3]. In addition, 1. Introduction decrease in the number of heating hours of 22-139h during the heating season (depending on the combination of weather as timber production, drinking water or welfare and nature conservation. Forests store large amounts of carbon [1, and 2] forests areplay an important forthetheother growing demand of a bio-based economy and they are a major renovation scenarios hand, function intercept increased for 7.8-12.7% pereffects decade (depending onfor the and thus a key considered). roleresource in the On global carbon cycle including mitigating climate change [3]. Insource addition, biodiversity. Thus a consistent pan-European gridded data set on the state of forest resources is essential for coupled scenarios). The values suggested could be used to modify the function parameters for the scenarios considered, and Today 40 % of the European land area is covered with forests managed for the provision of ecosystem services such forests are an important resource for the growing demand of a bio-based economy and they are a major source for improve accuracy heat demand estimations. as timber the production, drinking water or welfare and naturedata conservation. store largeresources amounts of [1,for 2] biodiversity. Thus a of consistent pan-European gridded set on theForests state of forest is carbon essential and thus play a key role in the global carbon cycle including mitigating climate change effects [3]. In addition, © 2017are TheanAuthors. Published by Elsevier forests important resource for the Ltd. growing demand of a bio-based economy and they are a major source for * Peer-review under responsibility of the91311 Scientific Committee of The 15th International Symposium on District Heating and Corresponding author. Tel.: +43 1 47654 biodiversity. Thus a consistent pan-European gridded data set on the state of forest resources is essential for Cooling. * E-mail address: [email protected]
Corresponding author. Tel.: +43 1 47654 91311 E-mail address: [email protected] 1876-6102 2017demand; The Authors. Published bychange Elsevier Ltd. Keywords:©Heat Forecast; Climate
Peer-review under responsibility of the scientific committee of the European Geosciences Union (EGU) General Assembly 2017 * © 2017author. The Authors. Published by Elsevier Ltd. Corresponding Tel.: +43 1 47654 91311 –1876-6102 Division Energy, Resources and the Environment (ERE). Peer-review [email protected] responsibility of the scientific committee of the European Geosciences Union (EGU) General Assembly 2017 E-mail address: – Division Energy, Resources and the Environment (ERE). 1876-6102©©2017 2017The TheAuthors. Authors.Published PublishedbybyElsevier ElsevierLtd. Ltd. 1876-6102
Peer-review under responsibility of of thethe Scientific Committee of The International Symposium Union on District Heating and Cooling. Peer-review responsibility scientific of15th the European Geosciences (EGU) General Assembly 2017 1876-6102 © under 2017 The Authors. Published by Elseviercommittee Ltd. –Peer-review Division Energy, Resources of and Environment (ERE). under responsibility thethe scientific committee of the European Geosciences Union (EGU) General Assembly 2017 – Division
Energy, Resources and the Environment (ERE). 10.1016/j.egypro.2017.08.052
2
Hubert Hasenauer et al. / Energy Procedia 125 (2017) 372–378 Hasenauer et al. / Energy Procedia 00 (2017) 000–000
373
researchers, policy makers and conservationists to study and understand the role of European forests for the global carbon cycle independent of political boundaries. In principle three different carbon monitoring methods are currently available: (i) forest inventory sampling based on repeated tree observations, (ii) flux tower observations recording the gas exchange between plants and atmosphere, and (iii) remote sensing information in combination with ecosystem modeling techniques providing continuous Net Primary Production estimates [4]. The purpose of this study is to use existing European data to develop a consistent pan-European data set for Net Primary Production (NPP), live tree carbon and volume per hectare, mean tree height and mean tree age by integrating remotely sensed satellite and harmonized forest inventory data from 13 different European countries. We address the conceptual challenges in comparing “space based” Moderate Resolution Imaging Spectroradiometer (MODIS) satellite driven Net Primary Production (NPP) versus terrestrial “ground based” productivity estimates using national forest inventory (NFI) data. We apply the biomass functions of each country as they are used for the national carbon reporting to estimate ground based NPP from repeated tree observations on each NFI plot. We link the terrestrial driven information with remotely sensed data to combine the advantages of the two approaches, e.g. the continuous and consistent coverage using remote sensing information, and capturing changes in the carbon allocation pattern due to forest management effects using the terrestrial data. Fig. 1 provides the conceptual outline of our study.
Fig. 1. Conceptual approach of the carbon estimation methods in combining top-down versus bottom-up data sources. The MOD17 algorithm provides Gross Primary Production (GPP) and Net Primary Production (NPP). Terrestrial NFI (National Forest Inventory) data provides tree carbon increment, which can be used to derive NPP by adding aboveground litter fall and belowground fine root turnover.
1.1. Consistent Net Primary productivity across Europe Net Primary Production (NPP) is an important ecological metric for studying forest ecosystems, their carbon sequestration, the potential supply of food or timber, and for quantifying the impacts of climate change [3]. For this study we obtain the original global MOD17 NPP estimates (called MODIS GLOB), which can be downloaded from http://www.ntsg.umt.edu/project/mod17#data-product. Since previous studies have indicated that MODIS driven NPP estimates can be substantially improved, if local climate data are used [4, 5], we rerun the original MOD17 algorithm with our newly down-scaled daily climate dataset across Europe [6]. This resulted in a new NPP dataset for Europe, called MODIS EURO [7]. The gridded MODIS EURO data are presented in Fig. 2.
374
Hubert Hasenauer et al. /Procedia Energy Procedia 125 (2017) 372–378 Hasenauer et al. / Energy 00 (2017) 000–000
3
Fig. 2. MODIS EURO NPP on 1-km resolution representing the periodic mean annual NPP for the period 2000-2012 using European daily climate data (available under ftp://palantir.boku.ac.at/Public/MODIS_EURO).
Next, we evaluated these two MODIS datasets with terrestrial driven NPP estimates from our national forest inventory data (NFI). The NFI data cover about 2 million trees with repeated observations from 196,434 inventory plots located in 13 European countries. We obtain the county specific estimation methods for calculating biomass as outlined in [8] and compared the three different NPP estimates (I) MODIS GLOB, (ii) MODIS EURO, (iii) NFI, as periodic mean annual NPP estimates by method for the years 2000 to 2012.
Fig. 3. Comparison of MOD17 NPP estimates: MODIS GLOB using the global climate data, MODIS EURO using 1-km daily European climate data) versus terrestrial driven ground NPP estimates using forest inventory from 13 European countries. The boxplots show the median, the 25th and 75th percentile, and the arithmetic mean (diamond). The whiskers extend to 1.5 of the interquartile range, values outside are indicated by circles. The number of observations (n) differs due to missing data for certain pixels.
4
Hubert Hasenauer et al. / Energy Procedia 125 (2017) 372–378 Hasenauer et al. / Energy Procedia 00 (2017) 000–000
375
We found that the global MODIS NPP data (MODIS GLOB) differ from NFI NPP estimates by 26 %, while MODIS EURO differs only by 7 % (Fig. 3). MODIS EURO agrees better with NFI NPP across scales (from continental, regional to country) and along gradients. The best agreement is evident for elevation, dominant tree species and tree height. One important task of our analysis was to assess the impact of forest management on the resulting NPP predictions. “Top down” remote sensing methods as implemented in MOD17 are based on the assumption of a full crown coverage to estimate NPP, while “bottom up” terrestrial NPP estimates are strongly driven by the number of trees on a given inventory plot. The number of trees on a given forest inventory plot may be expressed by competition measures such as the Stand Density Index (SDI) [9], which depends on forest management. Forests are managed for concentrating volume increment rates to fewer trees, e.g. change in the allocation pattern to enhance timber quality and stand stability. Thus we next calculated for each forest inventory plot the difference in NPP between MODIS EURO and NFI driven estimates (∆NPP) and plot the resulting differences versus SDI. Fig. 4 gives a typical example for Austria. For further details related to this analysis we refer to [7].
Fig. 4. NPP Difference (∆NPP) MODIS EURO minus NFI NPP versus Stand Density Index classes (SDI) for Austrian Forests. MODIS EURO overestimates NPP versus terrestrial driven NPP from NFI data at lower densities, while with increasing SDI ∆NPP converges toward zero. This indicates that higher SDI values mean maximum stand density which equals full crown cover. Lower SDI values suggest lower tree densities due to forest management an effect not explicitly addressed in the MODIS NPP data. Again the boxplots show the median, the 25th and 75th percentile, n is the number of observations by SDI class.
The results of Figs. 3 and 4 confirm that (i) using European daily climate data strongly improve the resulting MOD17 NPP estimates of European forests, and (ii) that MOD17 NPP estimates need to be corrected for potential stand density effects due to forest management to ensure unbiased and consistent NPP data for Europe [7].
1.2. Forest characteristics An important data source for policy makers and conservationists to study and understand European forests independent of political boundaries, is the availability of a consistent pan-European gridded dataset on the state of European forests. Although National Forest Inventory (NFI) data provide information on the characteristics of forests, including carbon content, volume, height, and age, such cross-national spatially consistent data do not exist for Europe. One option is to use remotely sensed data in combination with terrestrial data to derive spatially explicit
376
Hubert Hasenauer et al. / Energy Procedia 125 (2017) 372–378 Hasenauer et al. / Energy Procedia 00 (2017) 000–000
5
forest information across Europe, which are methodologically consistent and comparable. We link remote sensing and terrestrial data on a 0.133° grid optimized accuracy and resolution [10] to provide pan-European maps for four key forest variables representing the time period 2000-2010 [10]: (i) alive tree carbon, (ii) volume, (iii) tree height, and (iv) tree age. The results show distinct differences in the carbon storage of European forests due to biophysical limits and regional historic drivers such as forest management, which directly affect the carbon mitigation option of European forests. We used this data to assess the state of forest resources across Europe showing that mountainous regions have the highest stocking carbon and volume, central Europe has the tallest mean tree heights and Austria and Northern Scandinavia have the oldest forests in Europe expressed by the mean tree age. We performed a “leave-oneout” cross- validation where we estimate each cell using data independent from its location by country to provide a European accuracy and error estimate of our data. The same procedure was applied for a “country-wise” crossvalidation. In this validation, we estimate cell values using only cells that are not in the same country as the target cell. The validation results indicated that the error varies by forest characteristic but shows no bias for all tested forest characteristics [11].
Fig. 5. Pan-European map of volume (growing stock) using forest inventory data from 13 countries, consistent European NPP data and a gapfilling clustering algorithm (available under ftp://palantir.boku.ac.at/Public/ForestCharacteristics).
1.3. Climate limitations on potential forest structure Fig. 5 illustrates that growing stock of European forests varies. This variation may be due to forest management e.g. older forests exhibit higher stocking timber than younger forests but also due to the growing conditions. The regional growing conditions depend on site conditions (soil, temperature, precipitation, etc.) which strongly affect species distribution and forest productivity. Thus, we were interested how spatially and temporally climate conditions may limit forest structure throughout Europe [11]. We obtain our forest characteristics data and examine climate limitations on potential forest structure, relevant for assessing potential timber assortments or the suitability as wildlife habitat. First we derive a general forest structure response curves versus individual climate variables as
6
Hubert Hasenauer et al. / Energy Procedia 125 (2017) 372–378 Hasenauer et al. / Energy Procedia 00 (2017) 000–000
377
well as combinations of climate variables. Next we selected two climate response curves to assess how changes in climate may affect 8 specific forest types in Europe. The results suggest: (i) Boreal forests are limited by minimum temperature, while the Mediterranean forests are limited by maximum temperature. Temperate forests are limited by both climate variables temperature and precipitation (Fig. 6). As a result we can conclude that during the last 50 years, the change in climate has led to a decline in the potential average diameter at breast height by 5.0 %. A similar result is suggested for the potential basal area per hectare with an average decrease of about 6.5 %. While in the northern parts of Europe almost no changes were detectable, changes in potential tree growth are expected mainly in Southern Europe [12].
Fig. 6. Limitations by climate on potential forest basal area (BA) as it may depend on tmax the daily maximum temperature, tmin the daily minimum temperature, and prcp the daily precipitation rates.
2. Discussion and conclusions Linking remote sensing with terrestrial data permits novel insights in forest dynamics, their properties as well as their response to environment and climate across Europe. European forests exhibit an annual average carbon uptake of 577 gC/m²/year, which can be considered as the carbon sequestration potential. This number may be also seen as the potentially available resources for harvesting to address the increasing demand of a growing bio-economy (Figs. 2, 3, 4). We can also quantify the limits on potential forest structures and the current amount and properties of forest resources across Europe, independent of political boundaries and across gradients (Figs. 5, 6). This information may be important to optimize the size and location of bio refineries and/or power plants using biomass as a resource. Furthermore the information on forest structure including tree age and maximum tree dimensions, is important for nature conservation issues, such as the evaluation of the habitat suitability, etc. that can be reached under the local growing conditions of forests. Our studies demonstrate a general concept for providing consistent European forest data which can be easily improved if more terrestrial data are available (e.g. re-measurements of the terrestrial NFIs). Tree mortality due to competition, age, and environmental stress factors (e.g. drought, insects etc.) change
378
Hubert Hasenauer et al. / Energy Procedia 125 (2017) 372–378 Hasenauer et al. / Energy Procedia 00 (2017) 000–000
7
forest ecosystems and their productivity. With a recently finalized tree mortality study based on our data sets [13] we could examine the relationships between tree mortality and productivity within European forests.
Acknowledgements This work was conducted as part of the project “FORest management strategies to enhance the MITigation potential of European forests” (FORMIT). The research leading to these results has received funding from the European Union Seventh Framework Program under grant agreement n° 311970. We are grateful to the national forest research stations from Estonia, Finland, Sweden, Norway, Austria, Belgium (region Flanders), France, Germany, Czech Republic, Poland, Romania, Italy and Spain for providing the forest inventory data for our analysis. We thank the two anonymous revisers and the editor of this special issue, Prof.Dr. Michael Kühn, for helpful comments.
References [1] Gallaun, Heinz, Zanchi, Giuliana, Nabuurs, Gert-Jan, Hengeveld, Geerten, Schardt, Mathias, Verkerk, Pieter., (2010). EU-wide maps of growing stock and above-ground biomass in forests based on remote sensing and field measurements. Forest Ecology Management 260, 252– 261. doi:10.1016/j.foreco.2009.10.011 [2] Pan, Yude, Birdsey, Richard, Fang, Jingyun, Houghton, Richard, Kauppi, Pekka, Kurz, Werner, Phillips, Olivier, Lewis, Simon, Canadell, Josep, Ciais, Phillipe, Jackson, Robert, Pacala, Steven, McGuire, David, Piao, Shilong, Rautainen, Aapo, Sitch, Stephen, Hayes, Daniel, (2011). A large and persistent carbon sink in the World’s forests. Science. 333, 988–992. doi: 10.1126/science.1201609 [3] Jackson, Robert, and Baker, Justin (2010). Opportunities and Constraints for Forest Climate Mitigation. BioScience 60(9):698–707. doi:10.1525/bio.2010.60.9.7 [4] Hasenauer, Hubert, Petritsch, Richard, Zhao, Maosheng, Boisvenue, Celine, Running, Steve, (2012). Reconciling satellite with ground data to estimate forest productivity at national scales. Forest Ecology Management 276, 196–208. doi:10.1016/j.foreco.2012.03.022 [5] Neumann, Mathias, Zhao, Maosheng, Kindermann, Georg, Hasenauer, Hubert, (2015). Comparing MODIS Net Primary Production Estimates with Terrestrial National Forest Inventory Data in Austria. Remote Sensing 7, 3878–3906. doi:10.3390/rs70403878 [6] Moreno, Adam, and Hasenauer, Hubert, (2016) “Spatial downscaling of European climate data.” International Journal of Climatology. doi:10.1002/joc.4436 [7] Neumann, Mathias, Moreno, Adam, Thurnher, Christopher, Mues, Volker, Härkönen, Sanna, Mura, Matteo, Bouriaud, Olivier, Lang, Mait, Cardellini, Giuseppe, Thivolle-Cazat, Alain, Bronisz, Karol, Merganič, Jan, Alberdi, Iciar, Astrup, Rasmus, Mohren, Frits, Zhao, Maosheng., Hasenauer, Hubert, (2016). Creating a Regional MODIS Satellite-Driven Net Primary Production Dataset for European Forests. Remote Sensing 8, 1–18. doi:10.3390/rs8070554 [8] Neumann, Mathias, Moreno, Adam, Mues, Volker, Härkönen, Sanna, Mura, Matteo, Bouriaud, Olivier, Lang, Mait, Achten, Wouter., Thivolle-Cazat, Alain, Bronisz, Karol, Merganič, Jan, Decuyper, Mathieu, Alberdi, Iciar, Astrup, Rasmus, Mohren, Frits, Hasenauer, Hubert, (2016). Comparison of carbon estimation methods for European forests. Forest Ecology Management 361, 397–420. doi:10.1016/j.foreco.2015.11.016 [9] Reineke, L, (1933). Perfecting a stand-density index for even-aged forests. J. Agric. Res. 46, 627–638. [10] Moreno, Adam, Neumann, Mathias, Hasenauer, Hubert, (2016) Optimal resolution for linking remotely sensed and forest inventory data in Europe. Remote Sensing of Environment. 183, 109–119. doi:10.1016/j.rse.2016.05.021 [11] Moreno, Adam, Neumann, Mathias, Hasenauer, Hubert, (2017). Forest structures across Europe. Geoscience data Journal, in press, doi:10.1002/gdj3.45 [12] Moreno, Adam, Neumann, Mathias, Hasenauer, Hubert, (2017) Climate limitation on European Forest structures. Agriculture and Forest Meteorology. in review. [13] Neumann, Mathias, Mues, Volker, Moreno, Adam, Hasenauer, Hubert, Seidl, Rupert, (2017). Climate variability drives recent tree mortality in Europe. Global Change Biology, in press, doi:10.1111/gcb.13724
ScienceDirect ScienceDirect Energy Procedia 00 (2017) 000–000
Available online at www.sciencedirect.com Availableonline onlineatatwww.sciencedirect.com www.sciencedirect.com Available Energy Procedia 00 (2017) 000–000
ScienceDirect ScienceDirect ScienceDirect Energy Procedia 00 (2017) 000–000 Energy (2017) 000–000 372–378 EnergyProcedia Procedia125 00 (2017)
www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia
www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia
European Geosciences Union General Assembly 2017, EGU Division Energy, Resources & Environment, ERE EGU European Geosciences Union General Assembly 2017, Division Energy, Resources & Environment, ERE
Assessing theEuropean resources and mitigation potential of European forests Geosciences Union General Assembly 2017, EGU Theresources 15th International on District Heating Cooling forests Assessing the andSymposium mitigation potential ofand European Division Energy, Resources a,* a & Environment, b ERE c Hubert Hasenauer , Mathias Neumann , Adam Moreno , and Steve Running a,* a c Hubert Hasenauer ,feasibility Mathias Neumann , Adam , and82,Steve Running Assessing the of using theMoreno heat bdemand-outdoor University of Natural Resources and Life Science, Institute of Silviculture, Peter-Jordan-Str. 1180 Wien, Austria
Assessing the NASA resources and mitigation potential of European forests Ames Research Center, Mail Stop 204-14, Moffett Field, CA 94035-0001, USA University of Natural Resources and Life Science, Institute of Silviculture, Peter-Jordan-Str. 82, 1180 Wien, Austria temperature function for aof Forestry long-term district heat demand Numerical Terradynamic Simulation Group, College & Conservation, The University of Montana, Missoula, MTforecast 59812, USA a,* Research Center, Mail Stop 204-14, a b c NASA Ames Moffett Field, CA 94035-0001, USA a
b
c
a
b
c
Hubert Hasenauer , Mathias Neumann , Adam Moreno , and Steve Running a a b c I. University Andrićofa,b,c *, A. Pinaand , P. , J. Fournier B. Lacarrière , O. LeAustria Correc Natural Resources LifeFerrão Science, Institute of Silviculture,.,Peter-Jordan-Str. 82, 1180 Wien,
Numerical Terradynamic Simulation Group, College of Forestry & Conservation, The University of Montana, Missoula, MT 59812, USA a
b NASA Ames Research Center, Mail Stop 204-14, Moffett Field, CA 94035-0001, USA Abstract a IN+ Center for Innovation, Technology and Policy Research - Instituto Superior Técnico, Av. Rovisco Pais 1, 1049-001 Lisbon, Portugal c
Numerical Terradynamic Simulation Group, College of Forestry & Conservation, The University of Montana, Missoula, MT 59812, USA b
Veolia Recherche & Innovation, 291 Avenue Dreyfous Daniel, 78520 Limay, France Abstract c National and international carbon reporting systems require information on forest carbon This information can be derived Département Systèmes Énergétiques et Environnement - IMT Atlantique, 4 rue Alfredstocks. Kastler, 44300 Nantes, France from national forest inventory data and remote sensing. Here we present the conceptual challenges in assessing forest resources National and international carbon reporting systems information on estimates forest carbon stocks.from This13 information can be derived across Europe by combining MODIS satellite versusrequire terrestrial driven NPP calculated national forest inventory from national forest inventory data and remote sensing. Here we present the conceptual challenges in assessing forest resources Abstract (NFI) data covering 200.000 sampling plots. The results suggest that MODIS NPP predictions using local daily climate data and across Europe by desnsity combining MODIS satellite versus terrestrial driven estimates. NPP estimates calculated from 13leads national inventory addressing stand effects, provide realistic forest productivity Ignoring these effects to anforest overestimation Abstract (NFI) data covering 200.000 sampling plots. The results suggest that MODIS NPP predictions using local daily climate data and in the estimated carbon storage of reporting European systems forests derived satelliteon data. National and international carbon require from information forest carbon stocks. This information can be derived addressing stand desnsity effects, provide realistic forest productivity estimates. Ignoring these effects leads to an overestimation from national forest inventory data and remote sensing. Here we present the conceptual challenges in assessing forest resources in the estimated carbon storage ofMODIS; European forests derived from satellite District heating networks are commonly addressed in the literature asdata. one of the most effective solutions decreasing the Keywords: carbon; inventory; NPP; NFI, biomass; volume; MOD17; satellite; remote sensing; bioeconomy across Europe by forest combining MODIS satellite versus terrestrial driven NPP estimates calculated from 13 nationalfor forest inventory greenhouse gas emissions from the building sector. These systems require high investments which are returned through theand heat (NFI) data covering 200.000 sampling plots. The results suggest that MODIS NPP predictions using local daily climate data © 2017 The Authors. Published by Elsevier Keywords: carbon; forest inventory; MODIS; NPP;Ltd. NFI, and biomass; volume; MOD17; satellite; remote sensing; bioeconomy sales. Due to the changed climate conditions building renovation policies, heatthese demand in the future could decrease, addressing desnsity effects, provide realisticcommittee forest productivity estimates. Ignoring to an overestimation Peer-reviewstand under responsibility of the scientific of the European Geosciences Unioneffects (EGU)leads General Assembly prolonging the investment return period. in the estimated carbon storage of European forests derived from satellite data. 2017 – Division Energy, Resources and the Environment (ERE). main of this paper isbytoElsevier assess the feasibility of using the heat demand – outdoor temperature function for heat demand ©The 2017 Thescope Authors. Published Ltd. Keywords: forest inventory; MODIS; NPP; NFI, biomass; volume; MOD17; satellite; sensing; bioeconomy forecast.carbon; The district of Alvalade, in committee Lisbon (Portugal), was used as aremote case Union study.(EGU) The district consisted 2017 of 665 Peer-review under responsibility of thelocated scientific of the European Geosciences Generalis Assembly © 2017 The Authors. Published by Elsevier Ltd. buildings that vary in both construction period and typology. Three weather scenarios (low, medium, high) and three district – Division Energy, Resources and the Environment (ERE). Peer-review under responsibility of the scientific committee of the European Geosciences Union (EGU) General Assembly 2017 renovation scenarios were developed (shallow, intermediate, deep). To estimate the error, obtained heat demand values were – compared Division Energy, Resources the Environment (ERE). with results from aand dynamic heat demand model, previously developed and validated by the authors. 1. Introduction ©The 2017 The Authors. Published by Elsevier Ltd. results showed that when only weather change is considered, the margin of error could be acceptable for some applications Peer-review under responsibility of lower the scientific committee of the European (EGU)after General Assembly 2017 1. Introduction (the error in annual demand was than 20% for all weather scenariosGeosciences considered).Union However, introducing renovation 40 Energy, % of theResources European area is covered with forests managed for the provision of ecosystem services such –Today Division andland the Environment (ERE). scenarios, the error value increased up to 59.5% (depending on the weather and renovation scenarios combination considered).
asThe timber drinkingland water or ison welfare and nature Forests store large of services carbon [1, Today 40 production, %ofofslope the European area covered with forests managed for thetoprovision of amounts ecosystem such value coefficient increased average within theconservation. range of 3.8% up 8% per decade, that corresponds to 2] the and thus play a key role in the global carbon cycle including mitigating climate change effects [3]. In addition, 1. Introduction decrease in the number of heating hours of 22-139h during the heating season (depending on the combination of weather as timber production, drinking water or welfare and nature conservation. Forests store large amounts of carbon [1, and 2] forests areplay an important forthetheother growing demand of a bio-based economy and they are a major renovation scenarios hand, function intercept increased for 7.8-12.7% pereffects decade (depending onfor the and thus a key considered). roleresource in the On global carbon cycle including mitigating climate change [3]. Insource addition, biodiversity. Thus a consistent pan-European gridded data set on the state of forest resources is essential for coupled scenarios). The values suggested could be used to modify the function parameters for the scenarios considered, and Today 40 % of the European land area is covered with forests managed for the provision of ecosystem services such forests are an important resource for the growing demand of a bio-based economy and they are a major source for improve accuracy heat demand estimations. as timber the production, drinking water or welfare and naturedata conservation. store largeresources amounts of [1,for 2] biodiversity. Thus a of consistent pan-European gridded set on theForests state of forest is carbon essential and thus play a key role in the global carbon cycle including mitigating climate change effects [3]. In addition, © 2017are TheanAuthors. Published by Elsevier forests important resource for the Ltd. growing demand of a bio-based economy and they are a major source for * Peer-review under responsibility of the91311 Scientific Committee of The 15th International Symposium on District Heating and Corresponding author. Tel.: +43 1 47654 biodiversity. Thus a consistent pan-European gridded data set on the state of forest resources is essential for Cooling. * E-mail address: [email protected]
Corresponding author. Tel.: +43 1 47654 91311 E-mail address: [email protected] 1876-6102 2017demand; The Authors. Published bychange Elsevier Ltd. Keywords:©Heat Forecast; Climate
Peer-review under responsibility of the scientific committee of the European Geosciences Union (EGU) General Assembly 2017 * © 2017author. The Authors. Published by Elsevier Ltd. Corresponding Tel.: +43 1 47654 91311 –1876-6102 Division Energy, Resources and the Environment (ERE). Peer-review [email protected] responsibility of the scientific committee of the European Geosciences Union (EGU) General Assembly 2017 E-mail address: – Division Energy, Resources and the Environment (ERE). 1876-6102©©2017 2017The TheAuthors. Authors.Published PublishedbybyElsevier ElsevierLtd. Ltd. 1876-6102
Peer-review under responsibility of of thethe Scientific Committee of The International Symposium Union on District Heating and Cooling. Peer-review responsibility scientific of15th the European Geosciences (EGU) General Assembly 2017 1876-6102 © under 2017 The Authors. Published by Elseviercommittee Ltd. –Peer-review Division Energy, Resources of and Environment (ERE). under responsibility thethe scientific committee of the European Geosciences Union (EGU) General Assembly 2017 – Division
Energy, Resources and the Environment (ERE). 10.1016/j.egypro.2017.08.052
2
Hubert Hasenauer et al. / Energy Procedia 125 (2017) 372–378 Hasenauer et al. / Energy Procedia 00 (2017) 000–000
373
researchers, policy makers and conservationists to study and understand the role of European forests for the global carbon cycle independent of political boundaries. In principle three different carbon monitoring methods are currently available: (i) forest inventory sampling based on repeated tree observations, (ii) flux tower observations recording the gas exchange between plants and atmosphere, and (iii) remote sensing information in combination with ecosystem modeling techniques providing continuous Net Primary Production estimates [4]. The purpose of this study is to use existing European data to develop a consistent pan-European data set for Net Primary Production (NPP), live tree carbon and volume per hectare, mean tree height and mean tree age by integrating remotely sensed satellite and harmonized forest inventory data from 13 different European countries. We address the conceptual challenges in comparing “space based” Moderate Resolution Imaging Spectroradiometer (MODIS) satellite driven Net Primary Production (NPP) versus terrestrial “ground based” productivity estimates using national forest inventory (NFI) data. We apply the biomass functions of each country as they are used for the national carbon reporting to estimate ground based NPP from repeated tree observations on each NFI plot. We link the terrestrial driven information with remotely sensed data to combine the advantages of the two approaches, e.g. the continuous and consistent coverage using remote sensing information, and capturing changes in the carbon allocation pattern due to forest management effects using the terrestrial data. Fig. 1 provides the conceptual outline of our study.
Fig. 1. Conceptual approach of the carbon estimation methods in combining top-down versus bottom-up data sources. The MOD17 algorithm provides Gross Primary Production (GPP) and Net Primary Production (NPP). Terrestrial NFI (National Forest Inventory) data provides tree carbon increment, which can be used to derive NPP by adding aboveground litter fall and belowground fine root turnover.
1.1. Consistent Net Primary productivity across Europe Net Primary Production (NPP) is an important ecological metric for studying forest ecosystems, their carbon sequestration, the potential supply of food or timber, and for quantifying the impacts of climate change [3]. For this study we obtain the original global MOD17 NPP estimates (called MODIS GLOB), which can be downloaded from http://www.ntsg.umt.edu/project/mod17#data-product. Since previous studies have indicated that MODIS driven NPP estimates can be substantially improved, if local climate data are used [4, 5], we rerun the original MOD17 algorithm with our newly down-scaled daily climate dataset across Europe [6]. This resulted in a new NPP dataset for Europe, called MODIS EURO [7]. The gridded MODIS EURO data are presented in Fig. 2.
374
Hubert Hasenauer et al. /Procedia Energy Procedia 125 (2017) 372–378 Hasenauer et al. / Energy 00 (2017) 000–000
3
Fig. 2. MODIS EURO NPP on 1-km resolution representing the periodic mean annual NPP for the period 2000-2012 using European daily climate data (available under ftp://palantir.boku.ac.at/Public/MODIS_EURO).
Next, we evaluated these two MODIS datasets with terrestrial driven NPP estimates from our national forest inventory data (NFI). The NFI data cover about 2 million trees with repeated observations from 196,434 inventory plots located in 13 European countries. We obtain the county specific estimation methods for calculating biomass as outlined in [8] and compared the three different NPP estimates (I) MODIS GLOB, (ii) MODIS EURO, (iii) NFI, as periodic mean annual NPP estimates by method for the years 2000 to 2012.
Fig. 3. Comparison of MOD17 NPP estimates: MODIS GLOB using the global climate data, MODIS EURO using 1-km daily European climate data) versus terrestrial driven ground NPP estimates using forest inventory from 13 European countries. The boxplots show the median, the 25th and 75th percentile, and the arithmetic mean (diamond). The whiskers extend to 1.5 of the interquartile range, values outside are indicated by circles. The number of observations (n) differs due to missing data for certain pixels.
4
Hubert Hasenauer et al. / Energy Procedia 125 (2017) 372–378 Hasenauer et al. / Energy Procedia 00 (2017) 000–000
375
We found that the global MODIS NPP data (MODIS GLOB) differ from NFI NPP estimates by 26 %, while MODIS EURO differs only by 7 % (Fig. 3). MODIS EURO agrees better with NFI NPP across scales (from continental, regional to country) and along gradients. The best agreement is evident for elevation, dominant tree species and tree height. One important task of our analysis was to assess the impact of forest management on the resulting NPP predictions. “Top down” remote sensing methods as implemented in MOD17 are based on the assumption of a full crown coverage to estimate NPP, while “bottom up” terrestrial NPP estimates are strongly driven by the number of trees on a given inventory plot. The number of trees on a given forest inventory plot may be expressed by competition measures such as the Stand Density Index (SDI) [9], which depends on forest management. Forests are managed for concentrating volume increment rates to fewer trees, e.g. change in the allocation pattern to enhance timber quality and stand stability. Thus we next calculated for each forest inventory plot the difference in NPP between MODIS EURO and NFI driven estimates (∆NPP) and plot the resulting differences versus SDI. Fig. 4 gives a typical example for Austria. For further details related to this analysis we refer to [7].
Fig. 4. NPP Difference (∆NPP) MODIS EURO minus NFI NPP versus Stand Density Index classes (SDI) for Austrian Forests. MODIS EURO overestimates NPP versus terrestrial driven NPP from NFI data at lower densities, while with increasing SDI ∆NPP converges toward zero. This indicates that higher SDI values mean maximum stand density which equals full crown cover. Lower SDI values suggest lower tree densities due to forest management an effect not explicitly addressed in the MODIS NPP data. Again the boxplots show the median, the 25th and 75th percentile, n is the number of observations by SDI class.
The results of Figs. 3 and 4 confirm that (i) using European daily climate data strongly improve the resulting MOD17 NPP estimates of European forests, and (ii) that MOD17 NPP estimates need to be corrected for potential stand density effects due to forest management to ensure unbiased and consistent NPP data for Europe [7].
1.2. Forest characteristics An important data source for policy makers and conservationists to study and understand European forests independent of political boundaries, is the availability of a consistent pan-European gridded dataset on the state of European forests. Although National Forest Inventory (NFI) data provide information on the characteristics of forests, including carbon content, volume, height, and age, such cross-national spatially consistent data do not exist for Europe. One option is to use remotely sensed data in combination with terrestrial data to derive spatially explicit
376
Hubert Hasenauer et al. / Energy Procedia 125 (2017) 372–378 Hasenauer et al. / Energy Procedia 00 (2017) 000–000
5
forest information across Europe, which are methodologically consistent and comparable. We link remote sensing and terrestrial data on a 0.133° grid optimized accuracy and resolution [10] to provide pan-European maps for four key forest variables representing the time period 2000-2010 [10]: (i) alive tree carbon, (ii) volume, (iii) tree height, and (iv) tree age. The results show distinct differences in the carbon storage of European forests due to biophysical limits and regional historic drivers such as forest management, which directly affect the carbon mitigation option of European forests. We used this data to assess the state of forest resources across Europe showing that mountainous regions have the highest stocking carbon and volume, central Europe has the tallest mean tree heights and Austria and Northern Scandinavia have the oldest forests in Europe expressed by the mean tree age. We performed a “leave-oneout” cross- validation where we estimate each cell using data independent from its location by country to provide a European accuracy and error estimate of our data. The same procedure was applied for a “country-wise” crossvalidation. In this validation, we estimate cell values using only cells that are not in the same country as the target cell. The validation results indicated that the error varies by forest characteristic but shows no bias for all tested forest characteristics [11].
Fig. 5. Pan-European map of volume (growing stock) using forest inventory data from 13 countries, consistent European NPP data and a gapfilling clustering algorithm (available under ftp://palantir.boku.ac.at/Public/ForestCharacteristics).
1.3. Climate limitations on potential forest structure Fig. 5 illustrates that growing stock of European forests varies. This variation may be due to forest management e.g. older forests exhibit higher stocking timber than younger forests but also due to the growing conditions. The regional growing conditions depend on site conditions (soil, temperature, precipitation, etc.) which strongly affect species distribution and forest productivity. Thus, we were interested how spatially and temporally climate conditions may limit forest structure throughout Europe [11]. We obtain our forest characteristics data and examine climate limitations on potential forest structure, relevant for assessing potential timber assortments or the suitability as wildlife habitat. First we derive a general forest structure response curves versus individual climate variables as
6
Hubert Hasenauer et al. / Energy Procedia 125 (2017) 372–378 Hasenauer et al. / Energy Procedia 00 (2017) 000–000
377
well as combinations of climate variables. Next we selected two climate response curves to assess how changes in climate may affect 8 specific forest types in Europe. The results suggest: (i) Boreal forests are limited by minimum temperature, while the Mediterranean forests are limited by maximum temperature. Temperate forests are limited by both climate variables temperature and precipitation (Fig. 6). As a result we can conclude that during the last 50 years, the change in climate has led to a decline in the potential average diameter at breast height by 5.0 %. A similar result is suggested for the potential basal area per hectare with an average decrease of about 6.5 %. While in the northern parts of Europe almost no changes were detectable, changes in potential tree growth are expected mainly in Southern Europe [12].
Fig. 6. Limitations by climate on potential forest basal area (BA) as it may depend on tmax the daily maximum temperature, tmin the daily minimum temperature, and prcp the daily precipitation rates.
2. Discussion and conclusions Linking remote sensing with terrestrial data permits novel insights in forest dynamics, their properties as well as their response to environment and climate across Europe. European forests exhibit an annual average carbon uptake of 577 gC/m²/year, which can be considered as the carbon sequestration potential. This number may be also seen as the potentially available resources for harvesting to address the increasing demand of a growing bio-economy (Figs. 2, 3, 4). We can also quantify the limits on potential forest structures and the current amount and properties of forest resources across Europe, independent of political boundaries and across gradients (Figs. 5, 6). This information may be important to optimize the size and location of bio refineries and/or power plants using biomass as a resource. Furthermore the information on forest structure including tree age and maximum tree dimensions, is important for nature conservation issues, such as the evaluation of the habitat suitability, etc. that can be reached under the local growing conditions of forests. Our studies demonstrate a general concept for providing consistent European forest data which can be easily improved if more terrestrial data are available (e.g. re-measurements of the terrestrial NFIs). Tree mortality due to competition, age, and environmental stress factors (e.g. drought, insects etc.) change
378
Hubert Hasenauer et al. / Energy Procedia 125 (2017) 372–378 Hasenauer et al. / Energy Procedia 00 (2017) 000–000
7
forest ecosystems and their productivity. With a recently finalized tree mortality study based on our data sets [13] we could examine the relationships between tree mortality and productivity within European forests.
Acknowledgements This work was conducted as part of the project “FORest management strategies to enhance the MITigation potential of European forests” (FORMIT). The research leading to these results has received funding from the European Union Seventh Framework Program under grant agreement n° 311970. We are grateful to the national forest research stations from Estonia, Finland, Sweden, Norway, Austria, Belgium (region Flanders), France, Germany, Czech Republic, Poland, Romania, Italy and Spain for providing the forest inventory data for our analysis. We thank the two anonymous revisers and the editor of this special issue, Prof.Dr. Michael Kühn, for helpful comments.
References [1] Gallaun, Heinz, Zanchi, Giuliana, Nabuurs, Gert-Jan, Hengeveld, Geerten, Schardt, Mathias, Verkerk, Pieter., (2010). EU-wide maps of growing stock and above-ground biomass in forests based on remote sensing and field measurements. Forest Ecology Management 260, 252– 261. doi:10.1016/j.foreco.2009.10.011 [2] Pan, Yude, Birdsey, Richard, Fang, Jingyun, Houghton, Richard, Kauppi, Pekka, Kurz, Werner, Phillips, Olivier, Lewis, Simon, Canadell, Josep, Ciais, Phillipe, Jackson, Robert, Pacala, Steven, McGuire, David, Piao, Shilong, Rautainen, Aapo, Sitch, Stephen, Hayes, Daniel, (2011). A large and persistent carbon sink in the World’s forests. Science. 333, 988–992. doi: 10.1126/science.1201609 [3] Jackson, Robert, and Baker, Justin (2010). Opportunities and Constraints for Forest Climate Mitigation. BioScience 60(9):698–707. doi:10.1525/bio.2010.60.9.7 [4] Hasenauer, Hubert, Petritsch, Richard, Zhao, Maosheng, Boisvenue, Celine, Running, Steve, (2012). Reconciling satellite with ground data to estimate forest productivity at national scales. Forest Ecology Management 276, 196–208. doi:10.1016/j.foreco.2012.03.022 [5] Neumann, Mathias, Zhao, Maosheng, Kindermann, Georg, Hasenauer, Hubert, (2015). Comparing MODIS Net Primary Production Estimates with Terrestrial National Forest Inventory Data in Austria. Remote Sensing 7, 3878–3906. doi:10.3390/rs70403878 [6] Moreno, Adam, and Hasenauer, Hubert, (2016) “Spatial downscaling of European climate data.” International Journal of Climatology. doi:10.1002/joc.4436 [7] Neumann, Mathias, Moreno, Adam, Thurnher, Christopher, Mues, Volker, Härkönen, Sanna, Mura, Matteo, Bouriaud, Olivier, Lang, Mait, Cardellini, Giuseppe, Thivolle-Cazat, Alain, Bronisz, Karol, Merganič, Jan, Alberdi, Iciar, Astrup, Rasmus, Mohren, Frits, Zhao, Maosheng., Hasenauer, Hubert, (2016). Creating a Regional MODIS Satellite-Driven Net Primary Production Dataset for European Forests. Remote Sensing 8, 1–18. doi:10.3390/rs8070554 [8] Neumann, Mathias, Moreno, Adam, Mues, Volker, Härkönen, Sanna, Mura, Matteo, Bouriaud, Olivier, Lang, Mait, Achten, Wouter., Thivolle-Cazat, Alain, Bronisz, Karol, Merganič, Jan, Decuyper, Mathieu, Alberdi, Iciar, Astrup, Rasmus, Mohren, Frits, Hasenauer, Hubert, (2016). Comparison of carbon estimation methods for European forests. Forest Ecology Management 361, 397–420. doi:10.1016/j.foreco.2015.11.016 [9] Reineke, L, (1933). Perfecting a stand-density index for even-aged forests. J. Agric. Res. 46, 627–638. [10] Moreno, Adam, Neumann, Mathias, Hasenauer, Hubert, (2016) Optimal resolution for linking remotely sensed and forest inventory data in Europe. Remote Sensing of Environment. 183, 109–119. doi:10.1016/j.rse.2016.05.021 [11] Moreno, Adam, Neumann, Mathias, Hasenauer, Hubert, (2017). Forest structures across Europe. Geoscience data Journal, in press, doi:10.1002/gdj3.45 [12] Moreno, Adam, Neumann, Mathias, Hasenauer, Hubert, (2017) Climate limitation on European Forest structures. Agriculture and Forest Meteorology. in review. [13] Neumann, Mathias, Mues, Volker, Moreno, Adam, Hasenauer, Hubert, Seidl, Rupert, (2017). Climate variability drives recent tree mortality in Europe. Global Change Biology, in press, doi:10.1111/gcb.13724