Microclimatic differences and their influence on transpirational ... - TUM

6 downloads 1 Views 3MB Size Report
Mohammad A. Rahmana,∗, Astrid Moserb, Thomas Rötzerb, Stephan Pauleita ... Technische Universität München, Hans-Carl-von-Carlowitz-Platz 2,85354.

Agricultural and Forest Meteorology 232 (2017) 443–456

Contents lists available at ScienceDirect

Agricultural and Forest Meteorology journal homepage: www.elsevier.com/locate/agrformet

Microclimatic differences and their influence on transpirational cooling of Tilia cordata in two contrasting street canyons in Munich, Germany Mohammad A. Rahman a,∗ , Astrid Moser b , Thomas Rötzer b , Stephan Pauleit a a Strategic Landscape Planning and Management, School of Life Sciences, Weihenstephan, Technische Universität München, Emil-Ramann-Str. 6, 85354 Freising, Germany b Forest Growth and Yield Science, School of Life Sciences, Weihenstephan, Technische Universität München, Hans-Carl-von-Carlowitz-Platz 2,85354 Freising, Germany

a r t i c l e

i n f o

Article history: Received 5 April 2016 Received in revised form 20 September 2016 Accepted 3 October 2016 Keywords: Transpiration Street canyon Aspect ratio Sap flux density Urban trees Cooling effect

a b s t r a c t Urban trees can help to mitigate the urban heat island through evapotranspiration. However, growing conditions in cities are heterogeneous and micrometeorological conditions in street canyons can have a large impact on a tree’s transpiration. Here we investigated a common urban street tree species Tilia cordata of different ages and sizes, planted in two contrasting street canyons in a densely built neighbourhood within the centre of Munich, Germany: Bordeaux Platz, an open green square (OGS), and Pariser Platz, a circular paved square (CPS) with similar aspect ratio ≈ 0.5. The experiment was carried out during the exceptionally hot and dry summer of 2015. The open green square showed significantly higher wind speed and vapour pressure deficit but lower soil temperature, less negative soil moisture potential and also a smaller wind tunneling effect compared to the circular paved square. All these variables showed strong relationship with the sapflux density (Js) of trees grown at these two sites. They explain almost 78% of the variation in Js. On average Js values of the trees at Bordeaux Platz peaked to 0.30–0.35 ml cm−2 min−1 during the day compared to the peak values of 0.20–0.23 ml cm−2 min−1 for trees at Pariser Platz. Consequently, trees grown at the open green square showed higher peak cooling of 2.3 kW tree−1 or 82 w m−2 than at the circular paved square (1.9 kW tree−1 ). Interestingly, nocturnal cooling was higher for trees at the circular paved square with 18% of daytime transpiration compared to 2% for trees at the open green square. The study gave new insights into the cooling benefits of urban trees planted in open green squares compared to closed and paved squares and its temporal variation. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The cooling provided by urban trees is one of their important environmental benefits that has been extensively researched (McPherson et al., 1997; Taha, 1997; Gill et al., 2007; Ennos, 2010; Zhang et al., 2014) and is getting increasing attention due to the urban heat island (UHI) effect (Oliveira et al., 2011). Along with ongoing climate change accompanied by accumulated effect of heat and drought (Gill et al., 2013), urban trees appear to be a feasible option for climate change adaption and mitigation of urban areas (Shashua-Bar et al., 2010; Zhang et al., 2014). Trees cool outdoor spaces and buildings largely by shading them (Heisler, 1986; Zhang et al., 2013); at city scale they reduce the urban heat island by

∗ Corresponding author. E-mail address: [email protected] (M.A. Rahman). http://dx.doi.org/10.1016/j.agrformet.2016.10.006 0168-1923/© 2016 Elsevier B.V. All rights reserved.

evapotranspiration (Grimmond and Oke, 1999; Grimmond et al., 2010). Research on local cooling has involved both experimental studies and modelling and has provided consistent results. The effective temperature that people experience in open spaces, as measured by their physiologically equivalent temperature (PET) can be reduced by 7–15 ◦ C on hot days by tree shade (Matzarakis et al., 1999; Shashua-Bar et al., 2011; Armson et al., 2012, 2013; Lee et al., 2013). Similarly during hot summer days permanent tree shade can reduce surface temperature by 12–20 ◦ C (Gill et al., 2007; Armson et al., 2012) and air conditioning costs of buildings by 30–50% (Akbari et al., 1997; Millward et al., 2014). However, assessment of regional cooling benefits from urban forests is limited due to the lack of measured data on boundary layer cooling. The large amount of research that has compared the air temperatures within urban parks with built-up areas unfortunately confounds local and regional effects and cannot be meaningfully used to determine the overall effect of urban trees (Bowler et al., 2010). More

444

M.A. Rahman et al. / Agricultural and Forest Meteorology 232 (2017) 443–456

useful estimates of boundary layer cooling can be determined by integrating evapotranspirational water loss from trees grown in different growth conditions. Apart from biological factors such as tree size, leaf area index (LAI); microclimate factors such as radiation, vapour pressure deficit (VPD), wind speed and direction, air temperature; edaphic factors such as soil moisture content and soil temperature are the major drivers of tree transpiration (Wullschleger et al., 2000; Pataki and Oren, 2003; Tang et al., 2006; Tognetti et al., 2009). However, in cities micro-meteorological conditions can be significantly different within a very short distance due to the heterogeneity of built and open spaces (Offerle et al., 2007). Heat exchange is particularly varied in the complex geometry of street canyons (street length, width and buildings height) (Erell and Williamson, 2006; Emmanuel et al., 2007; Offerle et al., 2007) which might influence transpiration of trees, hence evapotranspirational cooling. Considering the practical difficulties in field experimentation (e.g. heterogeneity, high costs, vandalism) modelling approaches have been favoured generally using the Penman-Monteith equation (Allen et al., 1998) multiplied by plant coefficients. However, the use of this approach to estimate evapotranspiration in urban areas (Huang et al., 1987; Gill et al., 2013; Jacobs et al., 2015) might not be a true representation of the real situation since this kind of analysis assumes an ample supply of water, favourable conditions, and a large homogenous vegetation area (Allen et al., 1998). In a spatially heterogeneous city, with its often harsh growing conditions these assumptions are likely to be violated most of the time. For instance, having many separate areas of greenspace would be expected to increase evapotranspiration (Rahman et al., 2011) by the oasis effect, whereas drought would be expected to reduce it (Gill et al., 2013). Therefore, taking into account a small scale approach of continuous water loss from individual trees will lead to better appreciation of urban greenspaces. There has been no study yet to the best of our knowledge which investigated the continuous climatic effects of urban street canyon on the tree water relationship. Moreover, few studies have investigated the water use of different tree species (Pataki et al., 2011; Ballinas and Barradas, 2016) in relation to their biophysical control under an urban environment (Chen et al., 2011, 2012). Therefore, our current understanding of water use by urban trees is limited (Hagishima et al., 2007; Bush et al., 2008; Chu et al., 2009). In a highly sealed urban environment built surfaces usually store more energy and have higher surface temperatures compared to vegetated surfaces (Montague et al., 2000) which increase longwave radiation flux and sensible heat and ultimately increase air temperature but decrease relative humidity. While Kjelgren and Montague (1998) reported that Pyrus calleryana transpire 30% more water when growing in asphalt cut-outs than surrounded by turf, other studies have shown reduced stomatal conductance and hence less transpiration for trees grown over paved surfaces (Montague and Kjelgren, 2004). However, more experimental studies with mature trees in urban setting are necessary since most of our current understanding of urban tree water relations was derived from studies aimed to address other questions such as wood anatomy (Bush et al., 2008; Peters et al., 2010); irrigation schedules or horticultural management (Buhler et al., 2007; Nielsen et al., 2007), species suitability (Pataki et al., 2011), or seasonal and diurnal variability of tree transpiration along climatic variables and airborne pollutants (Wang et al., 2012). Similarly use of direct measurements such as eddy covariance ecosystem evapotranspiration measurements (Granier et al., 2000; Testi et al., 2004; Williams et al., 2004; Fisher et al., 2007) have been mostly done in non-urban conditions. There are only few studies such as Grimmond and Oke (1999) who used this technique on seven North American cities. They reported significant mid-day evapotranspiration rates up to 225 W m−2 during hot summer days for the urban environment.

Only a handful of studies (Rahman et al., 2011, 2013, 2014, 2015; Gillner et al., 2015; Konarska et al., 2016) have tried to directly estimate the whole water loss of urban trees to quantify the latent heat flux. All of these studies used gas analyser or porometer to estimate the stomatal conductivity and estimate the whole tree transpiration. However, porometers only take a snapshot of water loss and the results they give are difficult to scale up to the tree level which might lead to overestimations (Schulze et al., 1985). The easiest way would be to measure water loss using a lysimeter, but this method is only possible for grass or for small trees in pots. Therefore, the majority of the studies have reported water loss from trees using sap flow analysis since this method is seen as being more reliable for continuous monitoring of whole-tree water fluxes, at least in the medium term (Granier et al., 1996; Oren et al., 1996; Andrade et al., 1998; Kostner et al., 1998; Wullschleger et al., 1998). Among the sap flow measurement techniques, the Granier method has been particularly popular due to its simplicity, degree of accuracy, reliability, and low cost (Lu et al., 2004). Continuous measurements are needed to better understand the cooling efficiencies not only spatially but also temporally. Konarska et al. (2016) studied the effect of weather conditions and permeability on both daytime and nocturnal tree cooling by using a porometer. They reported significant night time transpiration. Night time transpiration has been reported by several other authors as well to lie in a range of between 5 and 15% among different species (Caird et al., 2007). This is very important particularly in the urban environment where night time temperatures can be significantly reduced due to the higher stability of the air and the shallow depth of the cooled air layer (the urban boundary layer during the night is limited to 100–300 m or less compared to 1 km during the day) (Oke, 1978). Moreover, wind speed is lowered due to reduced convection in urban areas at night (Oke, 1978) and a small amount of latent heat exchange can lead to a greater cooling effect. Considering these knowledge gaps this study aims to improve understanding of the biophysical control of the whole tree transpiration hence transpirational cooling spatially and temporally in the urban environment. Here we investigated a common urban street tree Tilia cordata, planted in two contrasting street canyons within the centre of a big German city of Munich. The specific research questions set for the experiment are i) Does street microclimate vary with street openness and if so, do they predictably influence the tree cooling potentiality? ii) Do the microclimate and plant response vary during the day and night? iii) How to assess the predictable variables for integrative thermal effects of street canyons in urban planning?

2. Material and methods 2.1. Study area The study site Munich (48◦ 8 N, 11◦ 35 E, at 520 m asl) is the capital and the largest city of the German state Bavaria and the third largest city in Germany after Berlin and Hamburg. It is a densely populated (4500/km2 ) city by German standards with a population size of 1.38 million (Bayerische Landesamt für Statistik, 2015). Close to the Alps, the climate in Munich is affected by its sheltered position and characterized by a warm temperate climate. The annual mean temperature is 9.1 ◦ C with a temperature range from −4 ◦ C (January) to 24 ◦ C (July) and with an annual precipitation of 959 mm, most occurring during summer with a maximum of 125 mm in July. The winter is drier with a minimum of 46 mm of precipitation in January (DWD, 2015). Munich’s architecture is characterized by only a few tall buildings higher than 100 m (Jochner et al., 2013); however, with frequent presence of deep street canyons (aspect ratio ∼ 2). Although a number of green

M.A. Rahman et al. / Agricultural and Forest Meteorology 232 (2017) 443–456

445

open areas can be found (Pauleit and Duhme, 2000) the city shows a strong UHI effect with monthly mean UHI intensity up to 6 ◦ C (Pongracz et al., 2010). Following a dedicated field campaign within the centre of Munich we selected two small squares with contrasting street canyon characteristics located within the eastern core area of the city: The Bordeaux Platz as an open green square (OGS), and the Pariser Platz as a circular paved square (CPS). Bordeaux Platz with avenue canyon and Pariser Platz with circular canyon but with similar aspect ratio ≈ 0.5. The neighbourhood is characterized by 3–4 storey perimeter blocks distributed in a regular configuration (Fig. 1). The main criteria involved were (1) street canyons with contrasting characteristics in terms of micro-meteorology, surface cover but within close proximity for the convenience of monitoring and experimentation (2) to have popular and healthy matured street trees of the same species present in a sufficient number and free from any visual decay or damage and (3) close to the city centre where the UHI effect is most pronounced and hence there is a particular need for tree shade by open space users during the summer heat. Among the two sites, the Bordeaux Platz an open green square (OGS) is characterized by an avenue plantation planted into lawn between two wide streets running from North to South and from South to North; the Pariser Platz, a circular paved square (CPS) is mostly paved square with 10 trees planted within. There was no irrigation, mowing of grass or pruning of trees at both of our sites during the entire experimental period.

spherical photos were also taken to estimate the sky view factor at the centre of both OGS and CPS sites.

Air temperature, air pressure, relative air humidity, precipitation, wind speed and direction were measured by installing two Vaisala Weather Transmitters WXT520 (EcoTech UmweltMeßsysteme GmbH, Bonn, Germany) at the two sites. At both sites the station was mounted on top of a 3.3 m street lamp post by a 3.5 m cross arm 2 m outward from the lamp to avoid influence of lamp and shade of the nearby trees and buildings. At OGS the weather station was placed in a straight line 16.8 m away from the first tree and at CPS 11 m away right at the centre of the square to capture differences in micro-meteorology as accurately as possible. At OGS on the same cross arm, a CMP3 pyranometer and a PQS1 Photosynthetically Active Radiation (PAR) sensor (Kipp & Zonen, Delft, The Netherlands) were installed to measure the global radiation and PAR respectively. All the data were recorded continuously at a 15-min resolution from 28th of July to 21st of October 2015 on enviLog remote data logger (EcoTech Umwelt- Meßsysteme GmbH, Bonn, Germany) attached to one of our sampled trees. The two sites were contrasting in terms of micro-meteorological differences especially in terms of wind speed and direction due to the shape of the canyons. CPS had lower wind speed, showed more channeling effect compared to the OGS (Fig. 2).

2.2. Tree selection and morphological measurements

2.4. Soil moisture potential and temperature measurements

T. cordata is widespread throughout Europe and the dominant street tree in Munich. It is characterized by a dense pyramidal or oval crown and commonly used in urban areas especially in sidewalks and in residential streets (Pauleit et al., 2002). At OGS there were four rows of T. cordata of approximately the same age. Two rows each were planted on the two sides of a pedestrian walkway on each side of the square (Table 1). We selected five trees of the most southern side and those were planted in the grass verges where approximately 50% of the rooting surface is underneath the grass verges while the other half is covered by the unpaved pedestrian walkway. At CPS 10 T. cordata were planted in small pits (4–4.5 m2 ). Hence, almost the entire rooting zone was underneath the pavement of the square or asphalt of the street. We selected five neighbouring trees in S, SW, W, NW and N direction. In both squares the influence of street canopies was masked by the shading effect of the buildings; however, at CPS the effect was more pronounced. Tree canopies at CPS were exposed to comparatively more direct solar radiation earlier in the day until 13:00 h and gradually had more shading compared to the tree canopies in OGS. Trees at the OGS were younger with significantly smaller diameter at breast height (DBH) than at the CPS although the total height and crown radius were not significantly different (Table 1). The average height of the branch-free trunk was about 5 m at the OGS and 4 m at the CPS. DBH was measured with a diameter measurement tape at a height of 1.3 m. Tree height was calculated using a Vertex Forestor. Crown radii were measured in eight inter cardinal directions (N, NE,..., NW) along the ground surface with a measuring tape from the centre of the trunk to the tip of the most remote downwardprojecting shoot and used to calculate crown projection area (CPA). LAI was derived from hemispherical photographs captured during the fully leafed phase (June–August) using a Nikon CoolpixP5100 camera with fisheye lens and Mid-OMount following Moser et al. (2015). The resulting hemispherical photos were analysed with the program WinSCANOPY (Régent Instruments Inc.). Among several methods we used LAI (2000 G)-Lin (Miller, 1967; Welles and Norman, 1991) as it resulted in the most reliable values. Hemi-

Soil matric potential and temperature at both the sites were measured using Tensiomark 1 (4244/1, range pF0-pF7) (EcoTech Umwelt-Meßsysteme GmbH, Bonn, Germany) installed vertically through soil profile to the depth of 30 cm. At OGS sampled trees were in a straight row and 7 m away from each other. The sensors were installed 3.5 m away from the main stem within the fringe of crown projection area, 3 sensors for the first tree and 2 for next 4 trees to have 3 replicates for each tree within the grass verges. However, at the CPS the same arrangement were not possible since the trees were planted into small open pits within the pavement. Here we installed two sensors for each tree at the furthest opening point from the main stem. All the sensors were installed in a place which was mostly shaded to minimize the effect of direct radiation on the soil surface.

2.3. Meteorological data collection

2.5. Sap flow measurements Tree transpiration was estimated from sap flux density (Js), measured continuously using thermal dissipation probes (TDPs) (Ecomatik, Dachau, Germany) introduced by Granier (1987). Pairs of 20-mm-long and 2.0-mm-diameter heating probes were inserted at a depth of 20 mm after removing the bark in the stem sapwood of 10 sampled trees of two plots. All sensors were inserted into the sapwood on the north side of the trunk at 4-4.5 m stem height from the ground to deter theft or vandalism because the two sites were public squares. Even after that, there was vandalism on both the plots for five occasions; hence, we lost some of our continuous measurement data especially from the CPS. The amount of sapwood area and its conducting role varies with species, tree age and environmental conditions (Cermak and Nadezhdina, 1998). Considering the heterogeneity of the two plots and the age of the sampled trees we installed two pairs of longer needles at a xylem depth of 20–40 and 40–60 mm with identical heating and sensing devices having the same diameter as those drilled for the outermost (0–20 mm depth) sensors. These longer sensors were placed radially between the northern and western sides of the trunk in proximity to the 20 mm sensor. We moved these two pairs of longer

446

M.A. Rahman et al. / Agricultural and Forest Meteorology 232 (2017) 443–456

Fig. 1. Plan view of Bordeaux Platz (OGS) and Pariser Platz (CPS) (up). Cross section of the street canyon (down) with sampled trees: (left) CPS; (right) OGS.

sensors in each plot to three individual trees in July, August and September to take into account the average Js along the sap wood radius. Within each pair of probes, the upper probe was constantly heated, whereas the lower probe was unheated and recorded the reference temperature of the wood. The electric current of the heating element was held constant at 0.12 A with a heating power of 0.2 W. The probes were encapsulated in aluminum tubes previously inserted into the stem and placed 15 cm apart to avoid thermal interference. All probes were covered with reflective foil and transparent plastic to minimize the influence of solar irradiance and air temperature. The temperature difference (T) between upper and

lower sensor probes was recorded every 30 s with a CR800 data logger (Campbell Scientific, U.K.) equipped with Campbell Logger Multiplexer, AM16/32B. Five-minute means were calculated from the 30-s readings and stored by the data logger. Temperature differences were converted to sap flux densities (Js; ml cm−2 min−1 ) based on Granier’s empirical calibration equation (eq. (1)) (Granier, 1987):

 TM − T 1.231

Js = 0.714

T

(1)

where TM is the maximum temperature difference when sap flow is assumed to be zero.

M.A. Rahman et al. / Agricultural and Forest Meteorology 232 (2017) 443–456

Wind speed OGS

2 1.5

180

1

90

0.5

0

Wind speed (m/s)

2.5

270

0 3:00

6:00

9:00

12:00

15:00

18:00

21:00 2.5

B

2

270

1.5

180

1

90

0.5

Wind speed (m/s)

360 Wind direction (o)

Wind direcon CPS

A

0:00

0

0 0:00 360

Wind direction (o)

Wind direcon OGS

3:00

6:00

9:00

12:00

15:00

18:00

21:00 2

C

270

1.5

180

1

90

0.5

0

Wind speed (m/s)

Wind direction (o)

360

Wind speed CPS

447

0 0:00

3:00

6:00

9:00

12:00 15:00 Time of the day

18:00

21:00

Fig. 2. Wind direction and speed of two canyons on three hot summer days (A) – on Julian day 219 (August 7), (B) – 220 (August 8) and (C) – 225 (August 13, 2015) at the two sites OGS (Bordeaux Platz) and CPS (Pariser Platz).

2.6. Sapwood area and total sap flow calculation Firstly each tree was cored to the heart wood at two opposing directions (N-S) and the sap wood depth was visually determined. Same tree core samples were used to estimate tree age and mean growth rate. Last year’s DBH increments were used to normalise all trees back to a DBH and associated sapwood area at an age of 40 years (Table 2). This enables direct comparison of water use amongst the trees at these two sites without the confounding effect of size and age differences. At both the sites the sap flux density significantly differs (ANOVA, P < 0.05) along the sap wood depth (Fig. 3). Averaged over the three months period, sap flux density for the sap wood depth of 20–40 mm was 27% and 15% higher compared to the outer most layers between 0 and 20 mm at the OGS and the CPS, respectively. On the other hand, Js for sap wood depth 40–60 mm was 48% and 18% lower compared to the outer most layers (0–20 mm) at the OGS and the CPS respectively. In order to consider these differences in terms of total sapflow calculation we divided the sapwood area into four sections (the average sap wood depth of five trees was 78.8 and 80.2 mm and estimated the total sap flow (SF) (ml tree−1 min−1 ) for the OGS (eq. (2)) and the CPS (eq. (3)). SF = Js ∗ SA/4 + Js ∗ 1.27 ∗ SA/4 + Js ∗ 0.52 ∗ SA/4 + Js ∗ 0.25 ∗ SA/4

(2)

SF = Js ∗ SA/4 + Js ∗ 1.15 ∗ SA/4 + Js ∗ 0.82 ∗ SA/4 + Js ∗ 0.65 ∗ SA/4

(3)

(ml cm−2

min−1 )

Here Js is the sap flux density at the sap wood depth of 0–20 mm, SA is the sap wood area (cm2 ). For the most inner part of the sap wood area (60–80 mm) we considered

the slope of the decrease (from 20–40 to 40–60 mm for example 1.27 − 0.52 = 0.75) and estimated the Js. Average SF were converted to daily values (i.e. multiplied by 60 × 24) and multiplied by the latent heat of vaporization LV which is 2.45 kJ g−1 to calculate the energy loss through transpiration (W tree−1 ) according to Eq. (4):

Energylosspertree = SF × L V × 60 × 24

(4)

2.7. Statistical analysis The software package R, version 3.2.1 (R Core Team, 2014) was used for statistical analysis. To investigate the difference between means both spatially and temporally we used Two Sample t-tests and for difference in Js in different depths we used one-way analysis of variance (ANOVA) with Tukey’s honest significant difference (HSD) test to identify the difference between the measured depths. In all the cases the means were reported as significant when p < 0.05. Pearson correlation coefficient was calculated to determine the association between different meteorological variables and paired scatter plot. Moreover, simple linear regression analyses were performed to determine the association between Js and each of the meteorological variables and finally, multiple linear model was developed based on the highest r2 values of individual independent variable.

448

M.A. Rahman et al. / Agricultural and Forest Meteorology 232 (2017) 443–456

Table 1 Average morphological characteristics of trees and degree of openness within the crown projection areas (CPA) at two sites: OGS (Bordeaux Platz) and CPS (Pariser Platz). Sites

CPA (m2 ± SE)

Degree of grass area within CPA (% ± SE)

Age (year ± SE)

Crown radius (m ± SE)

DBH (cm ± SE)

Height (m ± SE)

LAI in June (±SE)

LAI in August (±SE)

OGS CPS

67.12 ± 3.37 81.70 ± 3.97

57.03 ± 2.03 5.29 ± 0.33

40 ± 1.88 89 ± 8.24

4.58 ± 0.11 5.05 ± 0.13

29.18 ± 0.52 44.68 ± 1.27

15.12 ± 0.21 16.78 ± 0.29

2.43 ± 0.20 2.53 ± 0.15

2.39 ± 0.17 2.55 ± 0.20

Table 2 Biometric characteristics of the 10 sampled trees analysed for sap flux density–xylem depths relationship at comparable age at two sites: OGS (Bordeaux Platz) and CPS (Pariser Platz). OGS

Estimated age (years) DBH (cm) Estimated DBH at the age of 40 years (cm) Mean DBH increment (mm/year) Mean sap wood depth (cm) Mean sap wood area (SA) (cm2 ) Sap wood area at the age of 40 years (cm2 )

CPS

Tree 1

Tree 2

Tree 3

Tree 4

Tree 5

Tree 1

Tree 2

Tree 3

Tree 4

Tree 5

43 29.3 27.0 4.47 8.4 552 492

42 29.6 28.0 4.81 7.9 539 500

36 27.2 31.2 5.11 8 483 583

35 30.3 35 5.92 7.5 537 648

46 29.5 24.8 4.49 7.6 532 411

67 44.1 27.0 3.58 8.4 939 491

98 41.7 27.6 2.42 7.9 838 489

102 48.2 28.3 2.21 8.8 1089 540

112 47 26.6 1.69 7.5 931 511

68 42.4 30.9 3.18 7.5 822 552

Sap flux density at 0-20 mm

Js (ml/cm^2/min)

0.4

Sap flux density at 20-40 mm

Sap flux density at 40-60 mm

A

0.3 0.2 0.1 0 219

Js (ml/cm^2/min)

0.4

220

221

222

223

224

225

220

221

222 223 Julian day

224

225

B

0.3 0.2 0.1 0 219

Fig. 3. Sap flux density estimated from TDP sensors at three different depths over the warmest week (Julian days 219–225, August 07–13) at the OGS (Bordeaux Platz) (A) and the CPS (Pariser Platz) (B).

3. Results

of the measuring days indicate mostly sunny days with high vapour pressure deficit (VPD) and air temperature (AT) values.

3.1. Microclimatic conditions and soil moisture status 3.1.1. Global radiation and PAR The daily totals of global radiation (Itot ) reached the highest values on July 31 with up to 26 MJ m−2 day−1 and maximum PAR up to 2200 W m−2 in July 28 (Fig. 4). Both global radiation and PAR values gradually declined towards September and October 2015. However, higher values for both global radiation and PAR in most

3.1.2. Micro-climatic differences between two plots Average air temperature measured over the time was not significantly different (mean = 16.88 ◦ C) when comparing the CPS with the OGS (mean = 16.80 ◦ C) (Fig. 4 and Table 3). Unlike air temperature wind speed was almost double at the OGS (mean = 0.94 m/s) compared to the CPS (mean = 0.46 m/s) over the time measured (t = 83.88, df = 10860, p < 0.001) (Fig. 4 and Table 3). This effect of circular squares on wind speed was similar throughout the day (Table 4). Analogous to the wind speed, VPD values over the

M.A. Rahman et al. / Agricultural and Forest Meteorology 232 (2017) 443–456

449

Fig. 4. Daily totals of global radiation (Itot ) and daily maximum photosynthetically active radiation PARmax measured at the OGS (Bordeaux Platz); average rainfall and soil moisture potential (SMP) of both OGS and CPS (Pariser Platz); air temperature (AT), wind speed (WS), soil temperature (ST) and vapour pressure deficit (VPD) of two sites: OGS and CPS between July 28 to October 21, 2015 (missing values for CPS towards the end are due to vandalism).

Table 3 Mean difference between two sites: OGS (Bordeaux Platz) and CPS (Pariser Platz) in terms of AT (air temperature), WS (wind speed), VPD (vapour pressure deficit) and ST (soil temperature). Sites

AT

WS

VPD

ST

OGS CPS

16:80 ± 0.071 16:88 ± 0.071

0.94 ± 0.005*** 0.46 ± 0.002***

0.86 ± 0.009*** 0.77 ± 0.007***

15.90 ± 0.039*** 18.73 ± 0.055***

± Standard error of mean. *** Mean values varies significantly at p < 0.001.

450

M.A. Rahman et al. / Agricultural and Forest Meteorology 232 (2017) 443–456

Table 4 Temporal variations of sap flux density (Js) along with the micro[HYPHEN]meteorological variables (between day 209 and 294) wind speed (WS), air temperature (AT), vapour pressure deficit (VPD) and soil moisture potential (SM), soil temperature (ST) at two sites : OGS (Bordeaux Platz) and CPS (Pariser Platz). Site

Time

SM (MPa)

ST (◦ C)

WS (m/s)

AT (◦ C)

VPD (kPa)

Js (ml cm−2 min−1 )

OGS CPS t-value df p-value

00–03

0.64 0.71 −1.56 438 0.12

15.91 19.05 −8.31 406

Suggest Documents