How important is the choice of plant species for the green roof’s cooling effect? Comparison of the behavior of two CAM-facultative species and two C3 species under drought conditions

High temperature and substrate moisture deficit are common on green roofs during the summer season. Crassulacean acid metabolism (CAM) allows succulent species growing on green roofs to survive under drought, due to the day-time stomatal closure and malate production (Dark CO2 fixation). This reduces the transpiration (Tp), but, consequently, the cooling effect due to the transpiration itself. In this study we compared the transpiration rates of two CAM-facultative species (Sedum lydium Boiss and Sedum kamtschaticum Fisch.) and two C3 species of semi-xeric grasslands (Lotus corniculatus L. and Bromus erectus Huds.). CAM-facultative species switch between C3 and CAM photosynthesis to respond to environmental conditions, for instance lowering of substrate moisture content (SMC) (1, 2). Our aims were to understand the capacity of selected species to continue the evapo-transpiration process, as the substrate moisture content declines and to establish when the CAM behaviour occurs in CAM-facultative Sedum species. Experiment was carried out in the glasshouses of the DipSA Department, with 14°/26°C night/day temperature regime. Plants were grown in boxes (40 cm x 30 cm x 22 cm) filled with 10 kg of green roof substrate (Harpo/SEIC intensive substrate). In addition to plant species (three replicate boxes each) three evaporation control boxes, containing just bare substrate were used. At the start of the experiment all boxes were watered to field capacity (average SMC 26% weight/weight) and no supplementary water was provided until the SMC reached < 3.5% w/w. In L. corniculatus and B. erectus, this was on the day 15th from the start of the experiment, while for the two Sedum species it was on the day 30th. The following parameters were measured: daily evapo-transpiration rate and daily SMC, through gravimetric method; leaf Relative Water Content (RWC) three times a week for all the species (3); nocturnal malate accumulation through titration method with NaOH in Sedum leaves three times a week (4). SMC declined over the course of the experiment that ended when the SMC of the substrate reached extremely low values (around 2.5 – 3.5% w/w). Malate concentration in S. kamtschaticum leaves rose sharply when SMC was around 11% w/w, while in S. lydium two malate peaks were observed when SMC was respectively 14% w/w and 6.5% w/w. We hypothesized that the increase of the malate concentration in the Sedum leaves at 10 - 14% w/w of SMC could be associated with the switch from C3 to CAM metabolism (4). This hypothesis is supported by the evapo-transpiration rates significantly lower in the boxes of Sedum species, compared to ones with L. corniculatus and B. erectus as well as by the relatively high RWC values under drought (the minimum RWC value recorded for the Sedum species was the 75.8 %). The C3 species, instead, continued to transpire at the 10 - 13% w/w SMC values when both Sedum species changed their metabolism, but this caused a rapid decline of RWC (L. corniculatus reached the 38.5 % and B. erectus the 55 %). We conclude that until the 10 – 13% w/w of SMC, which support the Tp of C3 species considered, L. corniculatus and B. erectus have a potential to improve the cooling effect of a green roof than the two Sedum species used for the experimentation. Green roofs with L. corniculatus and B. erectus will however require more frequent irrigation than one with Sedum species which lose this cooling capacity sooner, but they show a greater resistance to drought. Further experiments will focus on the continuous monitoring of leaf and substrate temperature and linking this with the leaf stomatal conductance (gs) and the net assimilation rates (A). This will allow us to further elucidate the mechanisms underlying cooling provision and the actual differences in the extent of cooling between these species.