Helmholtz-Zentrum Geesthacht, 2017-03-29
http://www.hzg.de/030637/index_0030637.html.en

Impacts and costs: forestry, agriculture and ecosystems services

• Project baseline impacts of a global mean temperature increase of 2°C on forestry, agriculture and ecosystems services on pan-European level.

• Identify adaptation options for impact assessment of adaptation policies in WP10.

Flow diagram: modelling cluster & link with other WPs (click graphic to enlarge)

Tools

Five process-based ecosystem and generic vegetation models are used: EPIC (IIASA), LPJmL (PIK), CLM4.5 (JRC), DSSAT/SALUS (ENEA) and ORCHIDEE (CNRS-IPSL). In addition, G4M (IIASA) model is used to quantify impacts of +2C global warming on forests and forestry. GLOBIOM (IIASA) – a bottom-up global recursive dynamic partial equilibrium model of the agricultural, bio-energy and forestry sectors – is used to assess impacts and adaptations, and to generate monetary estimates of impact and adaptation costs. The WP7 bio-physical models were coupled with bias-corrected EURO-CORDEX dataset at 0.25° spatial resolution (LPJmL at 0.5°). The input variables from WP2-4 and model requirements are summarised in Table 1, and mandatory simulations for WP7 are listed in Table 2 (agreed at GA 2013, Rome). Model descriptions are provided in Deliverable 7.2.
Table 1. Inputs from EURO-CORDEX ensemble

Table 1. Inputs from EURO-CORDEX ensemble

Enlarge table

Impact2C WP7 Table2

Table 2. Mandatory EURO-CORDEX simulations

Enlarge table

1. Impact of 2C global warming on crop production

Preliminary simulations with EURO-CORDEX ensemble indicate that +2°C global warming will likely reduce yields of winter crops due to accelerated phenological development. For example, winter wheat yields will decrease by more than 0.5 t/ha in many western and central-European regions. However, changes in precipitation pattern during the growing season may lessen negative effect of temperature in some dry regions. In northern parts, increased growth rates of winter crops during winter and early spring may result in rising yields. Summer crops would be rather negatively affected in the Mediterranean region, but, in general, they will be more yielding at higher latitudes. For example, maize yields may increase by more than 1 t/ha in central and western Europe. Yields of summer crops are likely to increase by more than 70% for at north latitudes (above 55°N), even though the absolute yield may still be quite low. A positive impact by more than 20-40% was simulated for mid latitudes (between 45 and 55° N), with a particularly positive response at higher altitudes. The ensemble-median impacts on yields of selected crops are presented in Fig 1.

Fig. 1. Ensemble-median impact of +2C global warming on rainfed crop yields (in tons of dry matter/ha) calculated by the EPIC model.

1a) spring barley

1a) spring barley

1b) maize

1b) maize

1c) winter rape

1c) winter rape

1d) winter rye

1d) winter rye

1e) winter wheat

1e) winter wheat

1f) sunflower

1f) sunflower

1g) soya

1g) soya

1h) rice

1h) rice

2. Impact of 2C global warming on ecosystems

From the ecosystem perspective, the impacts of +2°C global warming on carbon in vegetation and ecosystems, soils, gross primary production (GPP), net primary production (NPP) and net of ecosystem exchange and production have been calculated with CLM4.5. For example, the results with KNMI_RACMO22E_EC-EARTH, RCP4.5 presented in Fig 2 show an overall increase in the amount of carbon fixed by GPP and NPP (an increase by around 10 to 40% in most of the grid cells in Europe) as well as the total fixed in ecosystems, but not in soils (Fig 2d). Soil carbon tends to decrease especially in boreal regions, while it is expected to increase in temperate climates. The baseline simulations show that the atmospheric carbon is stored mainly in temperate deciduous forests and its ecosystems over Central and Eastern Europe.

Fig. 2. Average change (in %) in (a) GPP, (b) NPP, (c) ecosystem carbon (ECSYSC), and (d) soil C at +2°C period relative to the baseline period (1971-2000) for KNMI_RACMO22E_EC-EARTH, RCP45 (calculated with the CLM4.5 model).

2a) GPP

2a) GPP

2b) NPP

2b) NPP

2c) ecosystem carbon (ECSYSC)

2c) ecosystem carbon (ECSYSC)

2d) soil C

2d) soil C

3. Impact of 2C global warming on selected ecosystem services

We analysed the impacts of +2°C global warming on selected indicators of ecosystem services, namely soil organic carbon (SOC), water storage (including growing season precipitation and ET, annual PET and ET, runoff, and irrigation water) and plant nutrition (plant available phosphates and nitrates).

Global warming is expected to significantly alter the carbon budget and C-related ecosystem services since climate is one of the main forces driving changes in soil carbon. Weather determine soil temperature and water as well as plant biomass and residues, and it all effects SOC stocks and dynamics. It is generally expected that rising temperature will accelerate the SOC decomposition in many European environments, leading to thinning of the SOC stocks in agricultural land (Smith et al., 2005). IMPACT2C ensemble-forced EPIC simulations indicate that SOC will likely be reduced in cropland of Europe, particularly in the Mediterranean region and countries of Eastern and Northern Europe, with a decline by more than 10% in many places (Fig 3a). We also demonstrate that conservative cropland management such as reduced tillage, no-till and crop residue management may counteract these negative effects of climate change (Deliverable 7.2).

LPJmL and EPIC simulations with EURO-CORDEX predict a decrease in ET over the crops’ growing periods in southern Europe (Fig 3b). This can be explained by simulated shorter growing season and lower yields, and also by lower soil moisture. Annual ET decreases less strongly and increases across northern Europe: this pattern basically follows the simulated changes in PET. Modelled runoff changes show more distributed patterns, with regions of positive change alternating with regions of negative change (Fig 3c). These runoff changes mainly mirror the scenario-specific changes in precipitation but also the general temperature increase. The adverse changes to water availability partly explain the adverse changes in crop yields reported above. As a corollary to changing (potential) ET and runoff, irrigation water use is simulated to change in many areas – with predominantly decreasing water use, due in particular to reduced growing periods but also due to aggravated water limitations that do not allow for further use (see WP6).

Nitrogen and phosphorus are pivotal for the functioning of bio-geochemical cycles and thus for long-run soil fertility service. The lack of N and P may lead to serious limitations in crop production and ecosystem services, such as it may prevent agro-ecosystems from acting as carbon sinks or realizing their climate-related potential to be enhanced. In intensified agricultural systems, N and P are being supplied through chemical fertilizers or manure. Assuming constant application rates over time, the total soil nitrates will likely decrease relative to present days in almost all cropland, and especially in Eastern Europe (Fig 3d). It mainly corresponds with an overall decline in SOC and increased leaching, especially under less-effective nutrition supply from crop management. Since N and P concentrations are strongly driven by fertilizer inputs, the pattern may significantly change depending on producers’ practices and adaptations.

Fig. 3. Ensemble-median impact (in %) on (a) soil organic carbon, (b) growing season evapotranspiration, (c) water runoff, and (d) total soil nitrates calculated by the EPIC model.

3a) soil organic carbon

3a) soil organic carbon

3b) growing season evapotranspiration

3b) growing season evapotranspiration

3c) water runoff

3c) water runoff

3d) total soil nitrates

3d) total soil nitrates

4. Uncertainties and robustness of the impact assessment

In addition to the impact assessment we also focus on uncertainties/robustness of impact signal propagated from the ensemble. Simulations with EURO-CORDEX data revealed large variability in simulated yields when analysed across climate models and projections, implying a serious source of uncertainty. The uncertainty is especially profound for rainfed systems in the Mediterranean region. As an example, animations in Fig 4a,b illustrate differences in simulated maize yields at 2C period and yield changes relative to the historical period for seven EURO-CORDEX scenarios. As proposed in WP5, the impact robustness is based on the agreement between the climate scenarios, meaning that if 66% of scenarios have all positive or all negative impacts on future yields, projected impacts demonstrate high level of robustness (Fig 4c,d).

Fig. 4. Animation of a) EPIC-simulated maize yields (in tDM/ha) at +2C period and b) yield change (in %) relative to the historical period of 1971-2000 as resulted from EURO-CORDEX ensemble; (c,d) ensemble-median “robust” impact (at 66%) of +2C warming on rainfed and irrigated maize.

4a) Click picture for animated view 4a) Click picture for animated view

4b) Click picture for animated view 4b) Click picture for animated view

4c)

4c)

4d)

4d)

5. Drought vulnerability and risk

Increased frequency of drought events under +2C warming will make crop production more vulnerable and will increase risk of yield losses. We used the Probability Risk Analysis (PRA, van Oijen et al. 2013) to estimate vulnerability of crops to extreme drought and to quantify expected risks of yield losses. The drought was defined using the Standardized Precipitation Evapotranspiration Index (SPEI, Vicente-Serrano et al., 2010) as SPEI values being less than –1. The monthly SPEI is based for each region on the local long-term frequency distribution of precipitation and potential ET and is therefore a localised measure of drought. For every grid cell the average SPEI over 1970-2100 is zero and values less than –1 are considered dry extremes. We averaged SPEI from April to September to better represent main growing season. For example, Fig 5 demonstrates high frequency of dry years in the Mediterranean region when CSC_REMO2009_MPI-ESM-LR, RCP4.5 scenario is considered.

Probability (P) of extreme drought was calculated for NUTS2 regions as the fraction of the 30 years (around 2C) with SPEI lower than –1. Vulnerability (V) of crop yields was calculated as the difference in median yields between normal and dry years, while the risk of yield losses equals V * P. With PRA we provide a quantitative definition (in t/ha) of yield vulnerability to drought events, and the risk of yield losses at NUTS2 level.

For CSC_REMO2009_MPI-ESM-LR, RCP4.5 scenario, for example, the drought vulnerability of rainfed maize is more than 2 t/ha in most of the Mediterranean regions, with the production risk of more than 0.5 t/ha (Fig 5).

Fig. 5. (a) Mean SPEI calculated for the +2C period, (b) probability of extreme drought at around +2C, (c) drought vulnerability of rainfed maize yields, and (d) risk of maize yield losses calculated for NUTS2 regions (CSC_REMO2009_MPI-ESM-LR, RCP4.5).

5a)

5a)

5b)

5b)

5c)

5c)

5d)

5d)

6. Cost assessment

Monetary estimates of impact and adaptation costs have been framed with GLOBIOM. The overall impacts of climate change on crop production was stratified by major market regions covering Europe in GLOBIOM, and converted into changes in total vegetal calorie supply from crop activities, if excluding adaptations other than changes in sowing and harvesting dates and adjustment in input level within boundaries of existing management systems (Fig 6). These impacts were reframed into a larger context, as European agriculture is to a large extent connected to international markets, which will be affected by local biophysical impacts on other regions. The relative gain of Europe is better highlighted once compared to other main world regions (Fig 7, HadGEM2-ES, RCP8.5): the aggregated impact is negative at global scale (-4%), and alongside with Europe only a few regions have positive regional impacts (Middle-East and Northern Africa MENA, Community of Independent States CIS, and Oceania OCE). Other regions face moderate (Latin and Central America LAM, Sub-Saharan Africa SSA, Eastern Asia EAS and Southern Asia SAS) to more dramatic (Northern America NAM, South-Eastern Asia SEA) negative overall impact on crop yields. This should to a large extent participate to shape adaptation decisions in Europe, as producers should overall face better market conditions.

Fig. 6. Definition of GLOBIOM regions, and overall aggregated biophysical impact of climate change in various European regions. (click picture to enlarge) Fig. 6. Definition of GLOBIOM regions, and overall aggregated biophysical impact of climate change in various European regions. (click picture to enlarge)

Fig. 7. Aggregated climate change impacts on crops in main regions of the world for +2C scenario (HadGEM2-ES, RCP 8.5) (click picture to enlarge) Fig. 7. Aggregated climate change impacts on crops in main regions of the world for +2C scenario (HadGEM2-ES, RCP 8.5) (click picture to enlarge)

7. Adaptation options in agricultural sector

Various adaptation mechanisms were simulated with GLOBIOM. Fig 8 decomposes at the European scale the various adaptation mechanisms leading to the overall increase in supply following adaptation to a +2°C climate change scenario. It includes the aggregated climate change impacts, supply-side adaptations (improved production systems, incl. irrigation and fertilization = MGMT, allocation of production systems and crop shifts = ALLO), and market-related changes (e.g., increased net export due to gains in competitive advantage).

Fig.8. Decomposition of adaptations and impacts of a +2°C climate change scenario (HadGEM2-ES, RCP 8.5) on European agricultural sector. (click picture to enlarge) Fig.8. Decomposition of adaptations and impacts of a +2°C climate change scenario (HadGEM2-ES, RCP 8.5) on European agricultural sector. (click picture to enlarge)

For impact cost estimates we rely on conceptual definitions developed in IMPACT2C (Deliverable 5.3, Methods for impact and cross sectorial assessment, socio-economic scenarios and cost calculation). For the agricultural sector, cost and impact estimates were directly driven from the GLOBIOM model outputs using a social welfare approach. GLOBIOM provides an estimate of the social welfare related to land-use activities, as the sum of surpluses of producers involved in the provision of related goods and consumers that consume them. The difference of this metric between two scenarios incorporates in particular two important information that usually vary in opposite directions: the way producers would value the overall changes (as a change in gross margin, i.e. producer surplus), and the way consumers would the value same changes (as a change in food expenditure compared to what they would be willing to pay for the same quantity, i.e. consumer surplus). Welfare is then defined as the sum of consumer and producer surplus for a given geographic scale. We also account for socio-economic changes between now and the +2°C time horizon, independently from the climate change (baseline scenario). Consequently, we define the monetary impact of climate change as the difference in European welfare gain from 2010 to 2030 (+2°C time period for this example) between the baseline scenario and the baseline with a +2°C climate change scenario after adaptation. The impacts before adaptation are evaluated for two different assumptions. First, we run a baseline simulation with climate change biophysical impacts, in which we constrained all supply-side variables to take the same values as they would have for the baseline without climate change. Second, we do the same while in addition constraining in a similar fashion the endogenous trade variables. Theses ‘no-adaptation’ scenarios describe a situation in which there is no adaptation of production systems through changes in management, localization, specialization etc… but we still allow trade (for the first scenario only) and demand to adjust.

At European scale, the welfare from land-use activities would increase by 70.78 billion US$2000 per year between 2010 and 2030 without climate change. For the +2°C scenario, this welfare gain would be of +71.34 billion US$2000 per year with adaptation, but of +68.82 billion US$2000 per year without supply-side adaptations, and +63.83 billion US$2000 per year in case both supply-side and trade adjustments do not occur. Climate has thus overall positive monetary aggregated impacts on land-use related sectors in Europe of +0.56 billion US$2000/year with adaptation, but a loss of 1.96 to 6.95 billion US$2000/year without adaptation. Fig. 9 presents the same numbers for the various European regions, normalized by number of inhabitants (and expressed in 1000.US$2000 per year per capita).

Fig.9. Measuring costs of climate change. (click picture to enlarge) Fig.9. Measuring costs of climate change. (click picture to enlarge)

References:

Smith, J., Smith, P., Wattenbach, M., Zaehle, S., Hiederer, R., Jones, R.J.A., Montanarella, L., Rounsevell, M.D.A., Reginster, I., Ewert, F., 2005. Projected changes in mineral soil carbon of European croplands and grasslands, 1990-2080. Glob. Change Biol. 11, 2141–2152. doi:10.1111/j.1365-2486.2005.001075.x

Van Oijen, M., Beer, C., Cramer, W., Rammig, A., Reichstein, M., Rolinski, S., and Soussana, J.-F., 2013. A novel probabilistic risk analysis to determine the vulnerability of ecosystems to extreme climatic events, Environ. Res. Lett., 8, 015032, 1–7.

Vicente-Serrano, S. M., Begueria, S., and Lopez-Moreno, J. I., 2010. A multiscalar drought index sensitive to global warming: the standardized precipitation evapotranspiration index, J. Clim., 23, 1696–1718.

Participants

IIASA (lead)
PIK, JRC, ENEA, CNRS-IPSL, PWA

Deliverables

[an error occurred while processing this directive]