Nothing to regret: Reconciling renewable energies with human wellbeing and nature in the German Energy Transition

Although the transition to energy supply through renewables (RE) is, in general, politically accepted in Germany, its progress is slowed by conflicting interests, primarily nature conservation and protesting residents. This study aims to find ways to solve these conflicts in Germany. To this end, the researchers developed a geospatial model that calculates RE potentials and vulnerabilities of nature and humans. Both data input and some evaluation standards are variables in the model. The outcomes are compared to an estimated total energy demand in 2050. Two ambitious scenarios (“no regret” and “compromise”) show that a maximum of 4% of the German territory is available to meet the energy demand. This demand can be met using PV in urban areas and wind in rural landscapes without significantly impairing nature's and people's wellbeing. Solar parks and other potentials not considered in the model are treated as a reserve, which can be included if the energy targets are not met under the assumed scenario conditions. Such reserves also provide flexibility for co‐determination in public participation.


| INTRODUCTION
The transition to energy supply through renewables is one of the central global challenges of the 21st century. It is a major component of national strategies for achieving the international climate change agreements (COP 21) and ensures sustainable energy supply without the use of fossil fuels. Globally, the share of renewable energies (RE) in total primary energy was 14% in 2015. 1 Still, climate change is not the only challenge we face. The loss of biodiversity appears to exceed the planetary boundaries, even more than greenhouse gas emissions, [2][3][4][5] and it requires an integrated governance approach. Although the necessity of the energy transition is undisputed, the concrete implementation faces many conflicts concerning the scarcity of land and the impacts of RE installations on human health and nature. On the one hand, strong incentive mechanisms have motivated the expansion of renewable energies in several countries and especially in Germany. On the other, citizens' protests have delayed the construction of power grids and wind turbines in the vicinity of residential areas and in recreational landscapes. These challenges require an approach that integrates the modeling of future RE development potential with the respective vulnerabilities of nature and humans and the possibility to downscale sustainable energy targets. Furthermore, a recent analysis of data-uncertainty revealed that the use of coarse spatial data strongly affects the calculations of available area at the national level 6,7 and thus the projected energy revenue. Consequently, models should use the most current and detailed data in calculations on the national level, especially if targets should be downscaled.
Many approaches to determine RE generation potential have been developed for different countries with different input parameters, target values for energy demand or time frames for the conversion of the energy system. One early approach at the regional scale aimed to integrate nature conservation and energy transition. It used a broad set of criteria that included different ecosystem services, which are used to calculate the potential for sustainable RE use. This model can help identify the tradeoffs between efficient energy generation and impairment of the environment. 8,9 As of yet, scenarios until 2030 for many countries primarily use the LUT model, developed at LUT University Finland. It aims to minimize the total system costs and employs regional data such as: power and heat demand, existing power and heat capacities, financial and technical parameters, as well as limits for installed capacity of all available technologies. 10 The calculations were made for large areas such as the Middle East and North Africa region (MENA region) 11 or Europe. 12 The model gives a good overview of the effective energy potential and its relation to demand. However, it offers only course spatial specificity and does not consider nature conservation restrictions in sufficient detail.
Furthermore, economic criteria that determine the total system costs drive the scenarios calculated for a Europe of 100% RE in 2050. 13 The spatial analyses used in these scenarios are based on wind and photovoltaic capacities in locations with optimized energy potential that are outside protected conservation areas. The potential is calculated on the basis of the European Reanalysis Interim (ERA-Interim) weather dataset and Corine Land Cover (CLC), and thus they remain spatially unprecise.
Although these modeling approaches provide good overviews, they cannot replace necessary national analyses that are required to achieve the national obligations under the Paris agreement. The LUT model was used for several national case studies that calculate, for example, the potential future developments in the energy system of Kazakhstan, 10 Turkey 14 and South Africa. 15 Othernational studies have calculated the energy potential using rather coarse input data 16,17 and considering only nature within protected areas. Furthermore, all studies assume presentday conditions and do not consider future developments in electricity demand, expected for technological innovations or land availability in the next 30 years. [18][19][20][21] Kienast et al 16 also work on the national level and also include selected landscape services similar to Palmas et al. 9 Their research also reveals the limits of studies that are economically focused by highlighting the conflicts and restraints at the local level of expanding renewable energies. Once again, their study assumes the existing generation technologies and land uses as the status quo and does not consider future development opportunities. For Germany, the existing studies on the national level face a similar situation with respect to spatial resolution 22 and limited scope.
In light of the described gaps in research, our study aims to determine whether Germany can produce enough RE on the national territory to meet the demand in 2050 without impairing humans or nature. Furthermore, our study allows for flexibility at the local level in shaping the energy landscape. This required an approach that permits us to estimate the future demand for renewable energy and calculate possible RE generation, while considering the evolving technologies as well as societal and natural restrictions. For this purpose, the energy demand in 2050 was projected and a spatial model was developed that considers both the potential RE production and the vulnerability of nature and humans. From the beginning, the research has been closely aligned with the planning process, as this facilitates the transfer of scientific results into practice and uses the language and approach of local actors.
The model uses scenarios to explore the potentials and limits of a sustainable energy transition. The scenarios and resulting policy and governance implications help decision makers identify implementation measures that can be used for a nature-friendly design of the energy transition process. In the initial development of the model, we concentrated on wind energy in rural areas and photovoltaic (PV) in urban areas because this selection reflected the current political stipulations and societal attitudes in Germany. 23,24 The model can illustrate the maximum potential of wind and solar energy, which in case of roof PV is not likely to be fully implemented. Ground mounted PV, in contrast, is perceived to be in competition with land uses such as agriculture, nature conservation, leisure and recreation. There is a fear that the displacement of such land uses could lead to land use intensification and impact biological diversity elsewhere. 25 However, its potential must be considered, for example, if electricity production by wind and roof PV cannot be fully exploited. In general, input variables related to political decisions that define the scenario path may vary and need adjusting to reflect societal preferences on different political levels.
Because many factors influence the implementation of energy projects, such as national and European funding programmes, we have not yet performed cost calculations. Rather, we focused initially on land availability, as it greatly influences the financing and realization options of RE. Thus, the determination of potential land areas forms the basis for planning the implementation of the energy transition. The projection of energy demand in 2050 was based on: the extrapolated population and economic development, an expected rate of electrification in the transportation sector, as well as assumptions for building renovations, and so on (see Table 1). The assumptions can be adapted to different future preconditions. Under these conditions, we projected the final energy demand for the year 2050 that uses both ambitious and moderate assumptions. The energy demand represents possible target values that are used to assess the modeling results. If the demand cannot be met, either more ambitious demand reductions are required or more RE must be produced to meet the 100% target.
The transition to a predominantly electric power supply system was assumed for all scenarios. This enables the coupling of all sectors of the energy system such as electricity, heating, mobility and industry. In addition to the basic electricity demand, we had to consider fluctuations in the daily and annual cycle of the energy production. These fluctuations lead to a high storage demand with corresponding conversion losses. Only some industrial processes and air transportation will continue torely on liquid raw materials. For this demand, combustibles must be provided from biomass (biogenic residues) and from "Power-to-X." Power-to-X refers to technologies that convert electricity generated from renewable sources into physical energy stores, energy carriers, and energyintensive chemical products.
Furthermore, the energy demand is based on the assumptions that heat pumps, which use ambient heat, are used to heat and cool households. The electricity demand for this technique is included in the calculation of the total energy demand.
2.2 | Modeling the sustainable electricity yield potential

| Overview of the modeling component
A spatial approach was chosen to model the sustainable electricity production because land availability is a decisive limiting factor and the protection of nature and as of humans occurs in a spatial context. The challenges and conflicts between nature conservation and energy production can be identified only on a site-specific basis by placing RE power generators in areas with a low vulnerability for humans and nature. 16 For this reason, the nationwide calculation of potentials for a supply of RE requires the utilization of spatial data. 18,20 Furthermore, downscaling in spatial planning requires spatially explicit information. Consequently, the model was built in a geographical information system (GIS). We used the GISsoftware ArcGIS 10.5 with the extensions "Geostatistical Analyst" and "Spatial Analyst".
The model includes the following components: Use of electric heat pumps (energy sources: electricity and ambient heat) 27 protection) and the determination of vulnerability classes with respect to RE generation (focus on wind and PV), in order to identify areas that should not be used for wind energy or only under specific conditions. This was achieved through: (i) literature-based definition of habitats and species sensitivity to power plants; (ii) expert-based definition of protection distances around other sensitive areas such as recreation landscapes; (iii) identification of areas of outstanding landscape beauty (according to 31 ). Furthermore, the objectives and targets of the German biodiversity strategy 32 were spatially specified in order to consider nature conservation development until 2050. 2. Area-specific calculation of electricity yield potential for onshore wind energy and PV on rooftops; 3. Overlay of the results of the previous analyses to identify low vulnerability areas for wind energy that are suitable for power plants and calculating the resulting overall potential electricity production for Germany.

| Scenario design and input variables
In order to avoid risks to humans and nature, the scenarios calculated for Germany consider only areas where wind energy is not expected to have relevant adverse effects. Four categories of spatial vulnerability were created, whereby high spatial vulnerable areas are excluded from the production of wind energy. In contrast, in areas of low spatial vulnerability the production of wind energy is compatible with humans and nature. One scenario, called "no regret", takes a particularly strict approach to a nature-and human-compatible expansion with renewables. The scenario "no regret" considers only areas that have a low spatial vulnerability to wind energy production, and thus, energy production in this scenario is compatible with human and nature needs without reservations. The second scenario, "compromise", also includes areas with medium vulnerability for allocating wind turbines that can be used with specific precautionary measures.
Details about the normative preconditions were defined in a transdisciplinary process with decision makers from federal administrations as well as NGOs. This process resulted in the decision that the project should focus on wind power generated outside urban areas and PV located inside urban areas. Thus, we have calculated the yield potentials for onshore wind energy and PV on roofs in a geoprocessing workflow, whereas the yield potentials of geothermal, hydropower and offshore wind energy were derived from existing studies (see Table 2). 41,42 A further project assumption was that the agricultural land presently used for food production should not be reduced in size.
Using input from experts, we chose a single type of wind turbines that represents the future technology for 2050as the technical input variable. Furthermore, experts also determined the assumptions about the efficiency of photovoltaics; making it possible to consider the further development of the technology up to the year 2050. Table 2 shows the technical layout of the two scenarios.
In addition to the technological developments until 2050, the research considers nature conservation objectives; this includes the present state of nature protection as well as the official objectives for 2050, that is, the "National Strategy on Biodiversity". 32 Potentially, these objectives could conflict with both the existing and future RE plants as well as the yet unimplemented requirements of the Federal Nature Conservation Act. For example, the rotor blades of wind turbines represent a vertical, movable element that protrudes into the airspace of birds and bats, resulting in collision casualties. The collision risk depends on the location, the height of the turbine, the rotor diameter as well as the number, density and arrangement of the turbines in the wind farm. 43 For the calculation of the individual spatial vulnerability classes (see Table 3), the vector data of the respective area categories were first transformed into a raster format with a resolution of 50 m × 50 m. We used only high resolution input data, such as the "Digital Basis Landscape Model" of the "Authoritative Topographic Cartographic Information System" (ATKIS Basis-DLM) from 2018, which has a positional accuracy of ±3 m and is based on the Topographic Maps 1:10000/1:25000. 50 Infrastructure, such as roads or railway lines, were depicted as lines in the spatial input data and buffered with average width before transformation into raster. Noise protection buffer zones were mapped around residential settlements that could be affected by noise emissions from future wind technology. Furthermore, the spatial vulnerability classes encompassed existing protected areas, other valuable ecosystems and habitats of sensitive species as well as areas with high visual landscape quality. 31 We also spatially specified future nature conservation goals related to the German biodiversity strategy and defined potential areas for nature conservation until 2050. 32 These extended nature reserves include military training areas, post-mining landscapes and forest or wilderness development areas that are excluded from wind park development. Areas with medium vulnerability include, for example, undissected low-traffic areas, morphological riparian zones, areas not strictly protected in the National Biotope Network or buffer zones and transition areas of biosphere reserves (see Table 3).
Locations of wind turbines were identified by applying a fishnet of rectangular cells with a resolution of 508 × 508 m to Germany. Wind turbines were then located in the center of fishnet cells in areas of the low spatial vulnerability (scenario "no regret") and the medium spatial vulnerability class (scenario "compromise"). The wind turbines were arranged with a safety distance of four times the rotor diameter (508 m, Table 2). 51 The on-site potential wind energy yield was calculated based on the average wind speed determined by the German Meteorological Service. This data depicts the wind speed at a height of 100 m above ground. 52 However, the wind speed was calculated at hub height of the wind turbine using the logarithmic wind profile, which can be taken to calculate the average wind speed at any given height. 22 Log wind profile is the wind speed at reference height in m/s, v(h 2 ) is the wind speed at new height m/s, hz 0 is the ground roughness in m. The calculation assumes a ground roughness of 0.10 m. In view of the spatial vulnerability, agricultural land was considered more suitable for wind turbines than structurally rich, forested areas or in built-up areas. 22,53 The framework for calculating the potential electricity yield uses the performance curve of the wind turbine 54 and the average on-site wind speed. Because the wind does not blow uniformly throughout, we calculated the relative distribution of the wind speeds. The relative wind speed frequency represents how consistently a wind speed occurs at a location. The relative frequency distribution was calculated using the Rayleigh distribution because the German Meteorological Service does not provide the Weibull parameters for all heights. K is assumed to be monotonically increasing in most cases. However, measurements have shown that this extrapolation is only valid to a limited extent. 55 From a height of 70-80 m the k-value decreases more strongly, making it difficult to determine k. Since the form parameter in Northern Europe is on average 2, we calculated the Rayleigh distribution using k = 2, which makes it possible to approximate the relative frequency distribution.
Rayleigh distribution (a k = 2 and k = 2): We used the usable roof areas and irradiation data (average values) to determine the potential electricity yield of PV. In addition to the existing buildings, we included the projected growth by 2050. To determine the usable roof area, we distinguished between residential buildings and industrial and commercial buildings. A saddle roof was assumed for residential buildings and a flat roof for industrial and commercial buildings. 22 However, not the entire surface of both pitched and flat roofs can be used for PV, because installations, ventilation shafts, chimneys or windows reduce the usable area. 56 We have assumed that on pitched roofs 60% 38,39 and on flat roofs 80% 22 of the surface can be used for modules with an efficiency of 30% (see Table 2).
The calculation of the usable roof area is based on the data set "house surrounds." The house surrounds describe georeferenced ring polygons of building floor plans and show the shape, location and extent of buildings. For flat roofs, the roof area corresponds to the ground area of the building. For pitched roofs, the roof area is calculated using the ground area and the pitch angle of the roof, 57 which is between 20 and 50 in Germany. 58 For the calculation of pitched roof areas, we assumed an average roof angle of 35 57 and included the south-facing half of the roof. We calculated the pitched roof potential (A) as follows: where a is the longer side of the ground plan, b is the shorter side of the ground plan.
A map of site-related PV production for Germany (showing kWh per year for the installation of 1 kWp) served as the basis for calculating the yield of PV on roofs, considering 30% efficiency and the potentially usable roof area. The efficiency of the PV systems was recalculated to determine the required area for the installation of 1 kWp, whereby 100% efficiency corresponds to 1 m 2 /kWp. On the basis of the roof area and information about the required area of the PV modules for the installation of 1 kWp, we determined how much kWp capacity can be installed on a building. Since pitched roofs rarely have an exact orientation to the south, the potential electricity yield was reduced depending on the orientation of the buildings.
Yield calculation for PV on roof areas Y ð Þ : where pra is the potential roof area, ur is the usable roof area factor, m is the m 2 required for 1 kWp, α is the roof angle, yr is the yield reduction due to orientation/slope. The higher deviation from the south orientation was, the greater the reduction in yield was. For example, 15% was deducted from the yield calculation of flat roofs because of the east-west orientation and 10 mounting of the modules.
In general, it should be added that nationwide only the used house surrounds are available as spatial data for the calculation of the potential area for rooftop PV. Therefore, we derived the assumptions that all house surrounds outside of industrial areas are covered with pitched roofs. This assumption in the model holds uncertainties, since flat roofs occur outside of industrial areas and industrial buildings can have pitched roofs.

| RESULTS
The main findings of the study are illustrated in two exemplary scenarios or snapshots of 2050. The snapshots clarify what is possible and point out alternative management variables.

| Energy demand in 2050
The projection shows that 1771 TWh/a are required to meet the German energy demand in 2050 under ambitious assumptions for demand reduction. The demand includes: 1227 TWh/a electricity (consisting of 818 TWh/a consumption plus an additional 50% of the electricity demand in order to compensate for storage and conversion losses) and 229 TWh/a ambient heat for use in electric heat pumps. In addition, other energy supply sectors must be decarbonized, for example raw materials for the chemical industry or fuels for air transportation. For these non-electrical sectors, a demand of 315 TWh/a was calculated, of which 60 TWh/a can be obtained from biogenic residues and waste materials. Furthermore, the electricity surplus that was calculated for the scenarios can also be used to meet the demand. The remaining 255 TWh/a can be produced in Powerto-X processes (ie, Power-to-Heat, Power-to-Gas, Powerto-Chemicals). However, the approx. 50% conversion losses of the Power-to-X may require 510 TWh/a of electricity in order to meet the 255 TWh/a demand. The demand for electricity thus increases to around 1737 TWh/a.
The projected total energy demand can be compared to the results of other studies that also simulate the German energy system for the year 2050. [59][60][61] They all assume a reduction in energy demand to around half of today's consumption, even though electricity demand will increase. This would require implementing intensive energy efficiency measures.

| Area potentials in 2050
In order to determine the area potentials for wind energy in the "no regret" scenario, the land area of Germany was classified into four classes that reflect the different levels of vulnerability of nature and human wellbeing to an exemplary standard wind turbine 2050 and its projected mechanical, visual and acoustic impact on the environment (see Figure 1).
The areas of the low vulnerability class cover a total of 5320 km 2 and thus represent 1.5% of Germany's land area (Figure 1). The areas of the medium vulnerability class cover 8303 km 2 , which represents 2.3% of the land area in Germany. The respective areas can be used for the production of wind power after a case-by-case examination of the local conditions. Thus, in areas with low or medium vulnerability the installation of wind turbines is, in principle, possible without causing unsolvable conflicts. For the PV potential, we calculated the possible energy yield using data about building perimeters. The potentially usable area of the pitched roofs comprises 2005 km 2 and 282 km 2 of the flat roofs are potentially usable for the installation of PV modules.

| Energy potential in the scenarios
In the first scenario "no regret," the calculation of the potential electricity yields included only areas with low vulnerability. Thus, on 5320 km 2 of land approximately 310 TWh/a electricity from onshore wind power can be produced. In addition, 900 TWh/a of solar power could be produced on the potential area of approximately 3000 km 2 of rooftops, facades and sealed traffic areas in Germany. Thus, a human-and nature-compatible total electricity yield potential of approx. 1 Wh/a can be produced sustainably. This would exceed the projected electricity demand in the target year 2050. PV electricity accounts for the largest share of the electricity mix at 64% (see Figure 2 and Table 4). We assume that this potential would even increase if spatial data were available nationwide to depict roof types. With a raise in the share of flat roofs, the potentially usable area would increase and thus the electricity yields. These uncertainties are integrated in the bar chart ( Figure 2) by a "phasing out" of yields. The electricity potential from onshore wind turbines has a share of 22%. Furthermore, we assumed 8% from offshore wind energy on the basis of present production as well as current contribution from hydropower and geothermal power.
In a second scenario "compromise", the areas with medium vulnerability to nature and human wellbeing were included in the calculation (an additional 8303 km 2 ). The production area for solar power remains unchanged. Under these assumptions, the potential electricity yield increases to nearly 1900 TWh/a (see Table 4).
Our model results show that the suitable area would be sufficient to cover the electricity demand of 818 TWh/a (+409 TWh/a electricity storage) in the year 2050 (see Figure 2). A surplus of 181 TWh/a (scenario "no regret") up to 664 TWh/a (scenario "compromise") of electricity could be produced. This surplus can cover the projected demand for non-electric uses, which adds up to 510 TWh/a including 50% conversion losses. Additionally, about 60 TWh/a would be available from residual organic substances as an energy source for nonelectric demand.
The land use for wind power plants is about 1.5% of Germany's land area in scenario "no regret" and 3.8% in scenario "compromise." The scenarios' assumptions still do not encompass all of the options to reduce or satisfy the energy demand in 2050 (see Figure 3). The potentials that have been excluded are: solar parks outside urban areas, a supplementary potential of off-shore wind and geothermal energy. Other factors not yet considered include changes in lifestyle, for example, in mobility, as well as using innovative technology that are still being tested, for example, technologies to avoid bird collisions or technological development in the field of energy efficiency. Furthermore, the import of sustainably generated energy from other countries was not considered in the defined limits of the scenarios. However, these unconsidered potentials can be seen as a "reserve" for the calculation of further scenarios. These untapped potentials for RE saving and generation also provide fall-back options. These options can be considered if poor implementation does not achieve the smart trajectory of the calculated scenarios or to provide public participation with additional options.
For political management of the energy transition, the model offers a flexible tool that brings demand and supply in line and helps to identify the necessary political action (Figure 3). If monitoring shows that implementation fails in one field then other variables can be introduced. For example, when the targets for site related energy generation are not achieved then demand can be decreased or other generation potentials can be activated. Furthermore, the model provides the information basis for defining local priorities that consider the trade-offs between impact on nature and energy yield. For instance, if areas that have medium vulnerability but very high potential electricity yields could be prioritized, then the overall space required for the energy transition would be minimized.
In summary, the study shows that human and nature compatible energy supply from 100% renewables are possible in 2050. The expansion of RE can be achieved when there is an optimal allocation of energy generation plants that follows the assumed, strict and very ambitious goals for landscape conservation and noise protection. Reserve potentials provide additional options for implementation, uncertainties and public participation. Energy source potential electricity yield in scenario "no regret" potential electricity yield in scenario "compromise" Although the implementation of the modeled scenarios requires determined political action, it would be a crucial step toward fulfilling the obligations of the Paris agreement without compromising biodiversity conservation or human wellbeing.
Some assumptions have been defined and implemented in the two scenarios. Others, such as the identified "reserves" (Figure 3), offer flexibility in further scenarios to react to changes in realization, while still bringing demand and supply into line. For example, although life style changes are variable, they should be included in the estimation of the renewable energy demand assigned to mobility (see Table 2). Similarly, energy supply could be achieved by using wind-only in areas that are not built-up, as shown in the scenario example, or by including ground mounted PV in rural areas. Our model shows the spatial changes that occur when assumptions are altered. Changes can lead to larger or smaller restricted areas for onshore wind. This, in turn, would change the yields from renewables.
In order to use the model to define the leeway for locally specified energy targets, a more detailed exploration of the model's uncertainties is necessary. The position accuracy and actuality of spatial input data is decisive for the reduction of uncertainties. For this reason, we used the "Basis Landscape Model" of the "Authoritative Topographic Cartographic Information System" (ATKIS Basis-DLM) for the display of infrastructures, settlements, agricultural land and forests. It is the land cover model with the highest positional accuracy in Germany; unfortunately, it is only available free of charge to research projects. The modeling assumptions include another source of uncertainties, which for example, stem from the assumption that all residential buildings have saddle roofs and all industrial and commercial buildings are covered by flat roofs. Furthermore, this input data could be improved. For instance, potentially 3D building models 62,63 could be used to specify actually usable roof area and facade potential. Much harder to quantify are the uncertainties related to the projection of the demand (demographic development or willingness to introduce new technologies).
In contrast to the ambitious assumptions of the scenarios, the present reality of energy governance in Germany is quite different. The established energy production paths and instruments are not yet sufficient to meet the challenges of a human-and nature-friendly energy supply.
Overall, the energy transition is stagnating in Germany, which has long considered itself a pioneer of energy transition. In the Annual Energy Transition Index, a worldwide ranking of the World Economic Forum 64 Germany has dropped to 20th place. Currently, a vast majority of the German population supports the energy transition process. 23,65 The "Social Sustainability Barometer of Energy system transformation" found that an average of 79% of interviewees view energy system transformation as a joint task to which everyone in society should contribute. 65 However, only 19% of those surveyed are "strongly in favor" of onshore wind energy, and a further 28% are "in favor" of it. However, 36% of the respondents supported the expansion of offshore wind energy.
This attitude is also evident in the implementation of the energy transition on the local level. In 2019, wind energy registered its lowest growth rate in more than 20 years. The newly installed plant capacity was 77% below the average new plant capacity of the previous five years. 66 The reasons for the slow expansion are diverse. In addition to the complex legal and economic framework conditions, citizen protests are becoming increasingly frequent, making it difficult to realize the energy potential. The conflicts arise on many levels: between the population and the decision-makers, but also within the community between those who benefit from the energy transition and the supposedly disadvantaged. The decisive factor for acceptance of energy system transformation seems to be the personal attitudes and participation options of individuals. 67 In many cases the overall goals of the energy system transition remain unclear, which makes the concrete implementation of energy projects at the local level more difficult. 68,69 As a result, the potential human and naturecompatible areas of the "no regret" scenario cannot be implemented without discussion and participatory processes.
In this context, the model can be used as a basis for public participation on the regional or local level when tailor-made targets for minimum RE generation are defined for individual political units. Such quantitative energy targets for local or regional political units, which in sum equal the national target, would enable flexible allocation and selection of energy mixes at the local level and avoid "not in my backyard" attitudes that jeopardize national targets. The definition of such energy targets would be based on the same national criteria and would clarify local responsibilities for the overall energy transition. At the same time, decision makers and citizens would be able to define their own implementation pathway within the context of the national target. Additionally, the model can help monitor how well regions achieve the national targets. Finally, the energy "reserves" offer additional options for an appropriate integration of local or regional preferences through local policies and public participation.
The changes to legislation and incentives that are required to make a difference are not extensive, but they are politically sensitive. Energy transition requires ambitious political frameworks that can be used to plan the energy system and the rapid implementation of energy technology innovations. This requires research about suitable combinations of legal, economic and communicative strategies and instruments.
A successful energy transition is also dependent on research and solutions from engineering science. For example, new developments in energy storage and grid expansion are immensely important. Future research can build on our approach by using spatial data to expand the possible applications of technical projections. Our scenarios are the first steps for the integrative modeling of energy systems that incorporate the spatial distribution of RE potentials as well as modeling consequences for grids and storage. The results of the study could also be used to calculate the cost of the transformation and to simulate transformation processes over time.

| CONCLUSIONS
Our study has demonstrated that conflicts between the goals of the transition to energy supply through renewables and the protection of humans and nature from impairments can be solved by intelligently allocating energy generation plants. Spatial modeling has proven to be a suitable approach to integrate both interests. Local targets for energy generation can be derived from the model and should, in sum total, achieve the national target. As long as targets for local energy saving and generation can be met, then local decision makers have flexibility concerning the RE mix and spatial allocation. Consequently, acceptance by the local population and decision making can be fostered without jeopardizing national targets through local "not in my backyard" attitudes. Finally, the model offers a flexible tool to bring demand and supply in line and to deduce the necessary political action at the national scale. If monitoring shows that implementation fails to decrease demand or increase RE generation in one option, then other variables can be introduced. The model helps decision makers to recognize which alternative management variables are at their disposal.
Further research work should thus focus in particular on analyzing different approaches of implementing the energy transition at local level.
The specific results of the scenarios cannot be directly transferred to other countries. However, the basic approach could be applied elsewhere, as the model can use country specific data. Nevertheless, further research is necessary in order to adapt the model to the specific data and implementation situation of different countries. Working out such concepts for all countries with RE surplus for export, will for instance avoid eco-dumping in a globalized market for hydrogen.
Germany is a densely populated country where different land uses are competing for the same land area. Therefore, the overall results are especially encouraging and informative for countries with similar pressures, where the approach can be instrumental for designing a sustainable energy policy.