3.3.1. Eastern Test Plots

The eastern part of the roof with the initial plantings of sedum cuttings developed differently at all times compared to the other the test areas on the west, north, and south sides with the turf mats. In 2011, all test plots achieved the minimum value of 60%. Table 10 shows the changes in the normal areas; Ulo values were lower than the Op values. The effect of fertilizer is greater on the very poor Ulo media compared to the better performing Op media.

**Table 10.** Greening with sedum cuttings, n = 9 years, from 2011 to 2020. The coverage and number of species of the test plots with fertilizer (Fert.) and without (Norm.) for the eastern exposure. Mean value, standard deviation, and kurtosis. The negative kurtosis value indicates that the distribution is characterized by weaker marginal areas than the normal distribution.


<sup>1</sup> Ulo = Ulopor media, Op = Optima media, Fert = Fertilizer, Norm = normal without fertilizer, East = Eastern section of the research roof.

The sedum coverage values increased on both media. The number of plant species decreased with the use of fertilizer.

The following Table 11 shows the statistics for the pairwise interpretation. In general, if the T value is not 0, the significance is high. On both media (Ulo, Op), the fertilizer significantly and positively impacted the vegetation cover. In addition, the decrease in the number of species was significant. The sedum coverage was only significantly enhanced on the Ulo test plot. On Op, the sedum coverage was not significantly more extensive because of the fertilizer but they bloomed better, as can be seen in Figure 2a.

**Table 11.** Greening with sedum cuttings, n=9 years, from 2011 to 2020 (df 8), coverage and number of species for the test plots with fertilizer (Fert.) and without (Norm.) of the eastern position. Pairwise T and the level 1‰ of significance differences between fertilized and non-fertilized plots in relation to cover values on Ulo. The minus symbols in relation to the Sedum cover; a reduction of these plants in relation to other plants.


<sup>1</sup> Ulo = Ulopor media, Op = Optima media, Fert = Fertilizer, Norm = normal without fertilizer, East = Eastern section of the research roof.

#### 3.3.2. Western test plots

The turf mat layout followed the same model as the eastern test plot. The basic mean values are shown in Table 12. The differences between the fertilized and non-fertilized sections on the Ulo test plots is less dramatic, with an 11% difference compared to the eastern test plot with the cuttings with a difference of 21%. This difference is significant, as can be seen in Table 13. The reduction in the number of species is also significant for both media, as Ulo dropped from 20 to 12 and Op dropped from 27 to 16, as both tables show. Again, the sedum coverage was significantly influenced by fertilizer usage but only on the Ulo media did it result in a significant change in the coverage value.

**Table 12.** Greening with vegetation mats,n=9 years, from 2011 to 2020. The coverage and number of species for the test plots with fertilizer (Fert.) and without (Norm.) for the western position. Mean value, standard deviation, and kurtosis.


<sup>1</sup> Ulo = Ulopor media, Op = Optima media, Fert = Fertilizer, Norm = normal without fertilizer, West = Western section of the research roof.

**Table 13.** Greening with turf mats, n = 9 years, from 2011 to 2020 (df 8), coverage and number of species for the test plots with fertilizer (Fert.) and without (Norm) of the western position. Pairwise *t*-test and the different level of significance. The significance indicate the effects of fertilizer on the cover values, negative sign, indicate reduction of the values over the years.


<sup>1</sup> Ulo = Ulopor, Op = Optima, Fert = Fertilizer, Norm = normal without fertilizer, West = Western section of the roof.

#### 3.3.3. Northern Test Plots

In the first years of this study, the northern test plots differed between the very shady areas near the building "North\_Shade" and those areas further away from the building with full sun, described here as "North Sun". This full-sun area differed from the southern test plots due to the additional heat reflected from the elevated building parts, with remarkably fewer grasses inside and a high dominance of the sedum cover. Table 14 shows that all north areas showed coverage values of above 90%. Consequently, the influence of the fertilizer was not significant in any case, as can be seen in Table 15 in the comparison of both Ulo Norm-Fertilizer pairs. Again, the reduction in the number of the plant species was significant for both growing media.

**Table 14.** Greening with vegetation mats, n = 9 years, from 2011 to 2020. Coverage and number of species for the test plots with fertilizer (Fert.) and without (Norm.) of the northern position; mean value, standard deviation, and kurtosis. Shade: in the shade of an elevated section of the roof.


<sup>1</sup> Ulo = Ulopor, Op = Optima, Fert = Fertilizer, Norm = normal without fertilizer, N = Northern section of the roof.

#### 3.3.4. Southern Test Plots

The southern section of the roof is much smaller than the northern section. It gets full sun all day plus the reflection from the metal façade of the elevated section of the building engineering area on the roof. This section of the roof also has an integrated air conditioning outlet that blows hot dry air over the southern green roof. The combination of these factors simulates extreme summer drying conditions for the green roof cover. The great performance of all the sedums growing here is remarkable. These succulents not only survive, but they also perform exceptionally well in these hot dry conditions. On the other hand, nearly all grasses are completely gone on this exposure.

The mean values of the southern plots are summarized in Table 16. With and without fertilizer, the coverage values are all above 90% and in general are the best of the test plots analyzed here. The sedum coverage values are above 70%, and with fertilizer they reach values above 80%. Again, the number of species decreased with the use of fertilizer. Table 17 presents the level of significance for these paired tests with the fertilizer test plots.

**Table 15.** Greening with turf mats, n = 9 years, from 2011 to 2020 (df 8), coverage and number of species for the test plots with fertilizer (Fert.) and without (Normal) of the northern position: North Shade: in the shade of an elevated section of the building. Sun\_N: North, outside this shade. Pairwise *t*-test and the level of significance. Example explanation line 1 of Table 15: The minus sign in category "means" symbols lower cover values of the first element of the pair. As example: Athough Ulo and Op were fertilized, over the time, Ulo has all the times lower cover values than Op on the highest level of 1‰ level.


<sup>1</sup> Ulo = Ulopor, Op = Optima, Fert = Fertilizer, Norm = normal without fertilizer, N = Northern section of the roof.

**Table 16.** Greening with vegetation mats, n = 9 years, from 2011 to 2020. Coverage and number of species, and the coverage of sedum only for the test plots with fertilizer (Fert.) and without (Norm.) for the southern position; mean value, standard deviation, and kurtosis.


<sup>1</sup> Ulo = Ulopor, Op = Optima, Fert = Fertilizer, Norm = normal without fertilizer, South = Southern section of the roof.

**Table 17.** Greening with turf mats, n=9 years, from 2011 to 2020 (df 8), coverage and number of species for the test plots with fertilizer (Fert.) and without (Norm.) of the southern position. Pairwise *t*-test and the level of significance. As explanation e.g. in line 1; also the south plots has a profit in cover value by the fertilizer, in this case on the 2‰-level. On the other hand, the fertilizer reduces the number of species. The Sedum cover is not significant influenced by the fertilizer.


<sup>1</sup> Ulo = Ulopor, Op = Optima, Fert = Fertilizer, Norm = normal without fertilizer, South = Southern section of the roof.

#### **4. Discussion**

Effective solutions are needed to stop the increase in average temperatures around the globe and to reach the targets in the Paris climate agreement in the next few years [30]. On the macro-scale, as explained in Fang et al. [31], for the landmass of China, drought, temperature, and global warming are significantly connected to the existing vegetation cover. Many studies have shown that cities around the world are in general drier than their surroundings, as can be seen from examples from China [32] and 70 cities in Europe [33]. In the search for solutions on a city-wide scale, it is apparent that any decentralized greenery improves the nearby living environment of its citizens [34]. Global warming is a multi-factorial issue with an energy–water nexus. Evapotranspiration is one main energetic factor, while CO2 is the accepted lead indicator from a political perspective to measure success in tackling climate change. However, not all factors can be explained by successful CO2 reduction. Green roofs in general and detailed technical solutions to improve the performance of green roofs as described in this publication are methods to adapt to and mitigate climate change. On a brighter note, energetic and chemical procedures are connected and need more holistic solutions.

The economic lockdowns during the COVID-19 pandemic were reported to lead to a 7% reduction in CO2 emissions in late 2020 and the start of 2021 [35]. Additional lockdowns are economically difficult and not considered acceptable for longer periods. More extensive and decentralized methods have to be considered. More CO2 fixation is required, and green roofs can contribute to this. Their impact is the step from no green roof to a green roof with relatively stable values of a total amount of about 6 kg/m2 as demonstrated in this survey. This value results from the total phytomass of shoots and roots of plants after a number of years of establishment, in this case 17 years. If we take this value as a baseline, extensive green roofs in Germany will result in about 51,000 t/CO2 fixation per year with about 8,500,000 m<sup>2</sup> of new green roofs every year. If this annual amount needs to be increased, higher vegetation values could be achieved with a thicker layer of growing media combined with annual fertilization. This survey showed annual growth rates on 30-cm media of about 100 to 600 g/m2 dry matter each year. However, this implied some maintenance of some form, such as mowing of the annual growth. Several earlier surveys measured the CO2 sequestration of green roofs. Our results correspond to the wide range of sequestration values observed, which usually only focus on the aboveground material, mostly sedum [36], with variation between 64 and 381 g/m<sup>2</sup> x years. Grasses in general perform better, as can be seen in a study from Japan [21] for *Cynodon dactylon* with sequestration of 2.5 kg CO2/m2\*year (fertilized and optimally irrigated roof modules) and *Sedum aizoon* (non-irrigated test plot of 1.2 kg CO2/m2\*year). The green roof rates are comparable to typical dry meadow habitats on the ground.

The CO2 fixation is an energetic procedure to tackle global warming. Finally, the size, quality, and distribution of the green roofs are important factors to have countable effects on the city scale [37].

Climate change is resulting in longer dry periods and lower humidity. Green roofs offer a range of benefits as demonstrated by several research projects over the last 20 years [38,39]. These reviews reveal that not all the results are comparable because the methods vary widely and most of the research was conducted over short time frames.

This 20-year survey confirmed that not all results could be achieved by a single roof project. What is important is the right selection of plant species, such as that shown in Table 18, which will help to achieve full vegetation coverage with many positive ecological effects.

**Table 18.** Preferred plant mixtures on the various roof spots of this survey.


Which plants are perfectly adapted to the expected climate changes? This survey was based on test plots with turf mats containing a generalized plant species mix. The test plots studied ranged from shade to extreme full sun, with some additional stress caused by high levels of solar reflection, air conditioner exhaust air outlets, and water stress.

An additional important factor is correctly choosing the right substrate. The test plot for this experiment shows that the type of growing medium is always critical for the biodiversity development over the years [40]. In most cases, highly biodiverse green roofs remain highly biodiverse over the years and support a wide range of species of birds and invertebrates [41]. As seen with this green roof research in Neubrandenburg, a significant quantity of mosses and lichens grow on very poor media. They can make up to a quarter of the total phytomass, resulting in up to 1.5 L/m2 \*day evapotranspiration [42]. On the one hand, the abundant presence of mosses and lichens characterizes areas with low pollution levels while also providing additional CO2 fixation. On the other hand, the massive growth of mosses and lichens displaces the typical or endangered higher plants that would otherwise be expected here [43].

Semi-intensive roof construction supports a wider range of plant species, such as dry-adapted prairie plants, with higher phytomass production and ultimately higher CO2 fixation [44]. The increasing number of urban agriculture sites may also be a solution for productive roof systems with massive phytomass production. Apart from the typical productive roof substrates, the competitiveness of hydroponic substrates, which have been used with increasing success in intensive green roofing, has improved in recent years [45]. The phytomass and species richness of typical extensive green roofs are comparable to dry meadows in a low mountain range [46]. Similar findings coming from North America, where more CO2 could be fixed by a green roof and intensive roof gardens can be the choice [47].

This survey also shows that over a time frame of 20 years, the number of plant species significantly decreases. If this is to be avoided, some maintenance work is helpful. Additional conclusions that can be drawn from this work are:



Table 19 compares the benefits of both methods, revealing that turf mats fulfill all requirements from the beginning, but they are more expensive. The effort for the initial maintenance is quite similar. The differences are apparent in the better CO2 fixation.

**Table 19.** Turf mat benefits compared to sedum cuttings as the primary vegetation cover, with results from this survey.


Fertilizer supports plant growth and plant performance in general [48] but not the local biodiversity. One single extra irrigation of 10 L/m<sup>2</sup> in early summer days had no significant effect on the plant performance of these dry-resistant herbs. In contrast to this, Du et al. [49] recommended emergency irrigation if shrubs are on green roofs. In times of increasing drought, the selection of the right plant species becomes an important issue. Lists with drought-tolerant plants have to be done on the regional levels, and the details in the microhabitat design support significant biodiversity [50]. Additionally, the focus on biological plant traits can help to develop an understanding regarding which species can be selected better in future times of climate change conditions [51]. In our survey, the importance of lichens of the genus *Cladonia* are highlighted as an important easy-to-care plant group and low-pollution condition, as they are in Neubrandenburg, a similar observation to a green roof working group in Canada [52]. Under the conditions of drought and climate change, prairie plants from North America become an alternative to be established in extensive green roofs under drought conditions for Central Europe [53]. Finally, the acceptance of successful growing weeds has to be accepted in natural concepts, and in many cases they can survive under extreme climatic conditions [54]. The need for local and regional biodiversity green roofs should be anchored in the related norms or guidelines in future [55].

Removing forest vegetation around the planet has significantly increased water shortages [56]. Heavier rainfall is also a consequence of the removal of vegetation in many places. This survey shows that minimal additional water has no influence on typical green roofs. Graduated tests of rain intensity and duration show the typical behavior, with extensive green roofs able to easily handle rain up to 30 L/m2 [57], with the right extra drainage elements below providing support [23]. Green roofs are effective in all climate zones around the world [58]. Green roofs work like a dewfall sink in cities. In comparison to typical bitumen roofs, green roofs capture the morning dew, which helps adapted plants to survive and is also a decentralized contribution against drought in cities [59].

Green roof technology has reached the international politics against global warming. They have become elements of the renovation strategy for building stocks and the EU biodiversity strategy 2030 [60]. The results presented here can contribute to ensuring more effective green roofs as part of the future wave of renovation [61] to successfully handle the coming climate challenges. Our current knowledge of green roofs reveals opportunities to construct new quality green spaces, which can be equivalent to vegetated spaces on the ground in terms of functionality and usability [62].

As recommendations for future research, this survey has shown that there is a need for more detailed research with more variations in growing media and structures to ensure that green roof spaces are more effective instruments of green infrastructure or systems for ecosystem services [63]. Edible cities are one of the new ecosystem services to which green roof spaces can contribute to in the future [64].

In addition, it is essential that methods between research institutions are aligned so that results can be more easily compared [65]. More long-term ecosystem studies can ensure better understanding of the variation in the integrated biological solutions. As a result of global warming and the adaptation of plants to growing on artificial urban surfaces, these systems react dynamically, unlike technical solutions that remain static.

#### **5. Conclusions**

This study modified the typical extensive green roof with a 10-cm media depth and low maintenance in the directions of deeper growing media with 30 cm. Further treatments were fertilization and irrigation. The aim was to enhance the CO2 fixation of the roof vegetation. However, what are the effects on vegetation cover and biodiversity? The summarized conclusions of these couple of years:




Green roofs can be established in higher quantities in all cities around the world. Although the best citizen/green roof values today exist in Stuttgart/Germany with 4 m2/citizen [13], and much more is possible if more regulation in legal plans is fixed.

Green roofs are one element in the strategies against global warming and can bring back a contact to man-made nature to citizens.

Typical extensive roof vegetation is well adapted to meet future climate change challenges, but detailed planning can help to develop green roofs to meet more specific aims, such as:





Green roofs are a small but visible factor bringing more evaporation surfaces into the anthropogenic urban desert and into the struggle against the drought.

**Author Contributions:** Conceptualization, M.K; methodology, M.K.; D.K.; Measurements M.K.; D.K.; validation, M.K.; D.K.; formal analysis, M.K.; D.K.; investigation, M.K.; D.K.; resources, M.K.; data curation, M.K.; D.K.; writing—original draft preparation, M.K.; D.K.; writing—review and editing, M.K.; D.K.; visualization, M.K.; D.K.; project administration, M.K.; All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding for the research; it is part of the Green roof lab of the University of Applied Sciences. We acknowledge support for the Article Processing Charge from the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation, 414051096) and the Open Access Publication Fund of the Hochschule Neubrandenburg (Neubrandenburg University of Applied Sciences).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The basic data sets of this research are archived in the University of Applied Sciences Neubrandenburg.

**Acknowledgments:** Special thanks to Marion Japp, Schwerin, for language improvement.

**Conflicts of Interest:** The authors declare no conflict of interest.
