Introduction

Human activities in rural areas have dramatically altered the nitrogen and water cycles over the last century. Wetland drainage and the widespread use of fertilisers have increased the amount of nutrients present in groundwater as well as in surface waters (Ripl 2003). The nutrient retention efficiency of natural and constructed wetlands used for agricultural purposes has been documented (Craft 1996; Fisher and Acreman 2004; Silvan et al. 2004; Hussain and Badola 2008; Powers et al. 2012). Wetlands are more efficient than other land cover types in performing these ‘services’, thanks to their ability to retain nutrients, stabilise nutrient cycles and enhance biodiversity within the landscape (Kao et al. 2003). In this study, we show that, in addition to these ecological functions, wetlands can strongly influence local and regional microclimates through evapotranspiration; an important process controlled by the interaction of several environmental and biological factors (Blanken et al. 1997; Wilson and Baldocchi 2000).

Solar energy warms the Earth at an average of 15 °C or 288 K. The amount of direct solar irradiance at the top of Earth’s atmosphere fluctuates through each calendar year from 1,412 W m−2 in early January to 1,321 W m−2 in early July, because of the Earth’s varying distance from the Sun on the elliptic trajectory (Geiger et al. 2003; Kopp et al. 2005). The amount of solar energy received at the earth surface varies by latitude and according to daily and seasonal pulses. The mean dissipation of solar energy in different months is shown on the NASA SSE web-site (http://eosweb.larc.nasa.gov/sse). The amount of energy received as incoming solar radiation within the temperate zone (Central Europe, Czech Republic) every year is about 1,150 kWh m−2, with slight variations depending on the latitude. The annual average in the centre-North of the USA reaches up to 1,365 kWh m−2, while in Florida (USA), it is 1,745 kWh m−2. The maximum irradiance (instantaneous flux) commonly lies between 800 and 1,000 W m−2 in the tropics and subtropics, which is similar to the irradiance occurring in temperate zones during the growing season. The amount of incoming energy differs significantly depending upon weather conditions. The difference between the maximum incoming solar radiation on a clear day can be an order of magnitude higher than the incoming radiation on an overcast day (Huryna and Pokorný 2010).

The largest part of the energy received during the growing season in most sites is converted into the latent heat of evapotranspiration. Functional properties and processes, such as soil moisture content, vegetative production and nutrient, and water cycling, are all influenced by evapotranspiration. The conversion of incoming energy into latent heat and sensible heat fluxes has a significant impact on local climate, as this process drives exchanges of energy and mass between the surface of continents and the atmosphere (Pielke et al. 1998). Makarieva and Gorshkov (2007) and Makarieva et al. (2013) described the fundamental role of forest evapotranspiration in the transport of water vapour from seas into continents and showed that the latent heat flux of evapotranspiration represents a process that directly links the energy and the water budgets of the landscape. The balance between the two plays a fundamental role in the cycling of water within terrestrial ecosystems and thus influences local climate patterns. Evapotranspiration is considered a driving force of landscape sustainability (Ripl 2003; Eiseltová et al. 2012). The water cycle is continuously driven by the Sun’s irradiation, and it plays a key role in the dissipation of solar energy and in the cycling of matter. Ecosystems, like other living systems, are dissipative structures in terms of non-equilibrium thermodynamic (Capra 1996; Schneider and Sagan 2005); they tend to release heat and to use solar energy for self-organisation.

When the Earth’s surface has no plant cover, or when there is not enough water for evapotranspiration, solar energy heats the land surface, which then transfers it to the air immediately above by generating sensible heat. Under high irradiance barren surfaces, such as agricultural fields after harvesting and urban tar-sealed surfaces become landscape ‘hot spots’, generating strong turbulence in the atmosphere above them (Kedziora 2004; Schneider and Sagan 2005; Kravčík et al. 2008).

The energy balance of a surface well supplied with water is presented in Fig. 1. The net radiation is the balance between incoming and outgoing short- and longwave radiations. In plants’ stands, the net radiation is largely dissipated through the evaporation of water from plants and soil, partly converted into sensible heat and partly transported into ground as the conductive soil heat flux. A minor part of net radiation is accumulated in biomass through photosynthesis.

Fig. 1
figure 1

Dissipation of solar energy in a plant stand well supplied with water. R s shortwave radiation, R l longwave radiation, α albedo, R n net radiation, H sensible heat, LE latent heat of evapotranspiration, G ground diffusion, J accumulation of heat in biomass, P energy consumption by photosynthesis

Full size image

The rate of evapotranspiration is determined by solar energy, air humidity, water availability, atmospheric pressure, as well as by relevant biological factors. The most important biological factor is the physiology of the plant species that cover the landscape surface, and their stage of development (Ryszkowski and Kedziora 1995).

A number of recent studies have focused on the dissipation of energy fluxes over several years (e.g. von Randow et al. 2004; Burba and Verma 2005; Liu et al. 2009); within a single year (Gu et al. 2005; Blanken et al. 2009); throughout the growing season (Ryszkowski and Kedziora 1987; Olejnik 1988; Wever et al. 2002; Kurc and Small 2004) and over several days (Olejnik et al. 2001; Pivec 2002; Eulenstein et al. 2005; Brom and Pokorný 2009; Rejšková et al. 2010). The seasonal heat balance of riparian buffer strips, meadows, agricultural fields (rape seed, beet and wheat) and bare soil was evaluated by Ryszkowski and Kedziora (1987) in the Turew region of Poland. They found that the net radiation ranged from 1.84 kWh m−2 d−1 (within the meadow) to 2.14 kWh m−2 d−1 (within riparian buffers) because of the differences in albedo. The energy consumed in evapotranspiration during the vegetative season ranged from 1.07 kWh m−2 d−1 for the bare soil to 1.88 kWh m−2 d−1 for the riparian buffer strips.

The roles of plants in the sequestration of carbon dioxide and in maintaining the oxygen concentration of the atmosphere are generally well understood. The photosynthetic exchange of CO2 and O2 is accompanied by a large release of transpired water: one molecule of oxygen is released for each molecule of utilised carbon dioxide, while several 100 molecules of water are evaporated (Schulze et al. 2002; You et al. 2009; Pokorný et al. 2010). It has been known from the 1950s that wetlands vegetation well supplied with water are able to convert several 100s (W m−2) of solar energy into evapotranspiration (Penman 1948; Monteith 1981). This direct effect of wetlands on climate has been overlooked, however, in the discussion on the impact of humans on climate changes. The indirect effect of wetlands on climate, i.e. sink/source of greenhouse gases (GHG), has been intensively studied and presented to decision makers and to the public (IPCC 2007). According to IPCC (2007), radiative forcing is the net change in the energy balance of the Earth system (stratospherically adjusted radiative flux change evaluated at the tropopause) in response to an increase in GHG in the atmosphere. It is equal to 1–3 W m−2 for the period from 1750 until present time. In the next 10 years, the radiative forcing is expected to increase by 0.2 W m−2.

The aim of our study is to show the direct effect of wetlands and their vegetation cover upon regional climate through the dissipation of the incoming solar radiation and its relation to the water cycle. Our study is based upon time-intensive quantitative measurements of dissipated solar energy, conducted over a whole year, comparing a wet meadow, a wet pasture, an open arable field and a concrete surface.

Methods and site description

Throughout the text, we will use the following terms, according to Parker (2002):

  • ‘Energy flux’ is a vector quantity component of which perpendicular to any surface equals the energy transported across that surface by some medium per unit area per unit time.

  • ‘Energy dissipation’ is any loss of energy, generally by conversion into heat (as well as into latent heat); its quantification consists in establishing the rate at which this loss occurs.

  • ‘Heat balance’ is the equilibrium which exists on average between the radiation received by the earth and atmosphere from the sun and that emitted by the earth and atmosphere.

The study was undertaken in an alluvial lowland within the Třeboň Biosphere Reserve in southern Bohemia, (TBBR) (48°49′ to 49°20′N; 14°39′ to 15°00′E) near the Czech–Austrian border. The flat bottom of the basin is at 410–470 m a.s.l., with an undulating margin reaching 550 m. The Třeboň Basin, the southernmost portion of the large Vltava catchment, belongs to a moderately warm region of the Central European temperate zone, with an annual mean temperature of 8 °C and a mean annual precipitation of 650 mm. The TBBR (700 km2 in total), has about 500 fishponds cumulative surface of which reaches 7,500 ha. Forestry, agriculture and fisheries have always been the main activities (Květ et al. 2002).

Four experimental sites were selected within a distance of between 2 and 4 km from each other:

  1. (1)

    A wet meadow (ca. 500 ha) within the Rožmberk fishpond (480 ha) catchment. Dominant species were high sedges (Carex acuta, Carex vesicaria), Calamagrostis canescens, Phalaris arundinacea, Urtica dioica. The area surrounding the meteorological station is not managed; tall vegetation is regularly cut only in the close vicinity of the meteorological station. The drier parts of the wet meadow are mown once a year. The wet meadow is an epilittoral (upper littoral) of Rožmberk fishpond which is inundated during an exceptionally high water level, mostly in spring or any rainy period. In the flat Třeboň Basin, such an epilottoral inundated area may be even larger than the fishpond area at standard water level.

  2. (2)

    A pasture with a relatively high groundwater level, characterised by dominant herbs such as Alopecurus pratensis, Ranunculus repens, Phleum pratense, Dactylis glomerata, Bellis perennis, Poa palustris, Trifolium repens, Taraxacum sect, and Veronica chamaedrys, Ruderalia.

  3. (3)

    A winter barley field (22 ha).

  4. (4)

    An artificial concrete surface (400 m2) located within the area of the Wastewater Treatment Plant of the city of Třeboň.

All four sites were exposed at similar angles to the sun rays, as the surface of the TBBR is flat. The dataset, used for the estimation of energy fluxes, evapotranspiration and micrometeorological conditions, was obtained from automatic meteorological stations M4016 placed at each research location. The details of measurement, equipment and sensors’ accuracy are summarised in Table 1.

Table 1 Details of in situ measurements, equipment and sensors’ accuracy
Full size table

The meteorological data were recorded during a whole year at 10-minute intervals. Complete series of data from the main part of vegetation season, when evapotranspiration rates were high (from 1 May to 31 August 2008), were used for the analysis. The data were arranged into the following three groups according to the amount of total incoming daily shortwave solar irradiance: overcast (0–3,000 Wh m−2), cloudy (3,000–6,000 Wh m−2) and clear (over 6,000 Wh m−2). The numbers of overcast, cloudy and clear days are presented in Table 2. Data collected on clear days were used to estimate the energy balance to evaluate the ability of the different land covers in dissipating solar energy and in dampening temperature extremes.

Table 2 Overcast, cloudy and clear days between 1 May and 31 August were used in the assessment
Full size table

We estimated energy used for evapotranspiration, vapour pressure deficit, aerodynamic and surface resistance to investigate the specific physical conditions prevailing at the different sites that could have influenced the monitored parameters. Nine days (three in a row) were selected for the analysis—at the beginning, in the middle, and at the end of the study period.

Albedo was computed as a ratio between reflected (Rs↑) and incoming (Rs↓) shortwave radiations:

$$\alpha = \frac{{R_{s \uparrow } }}{{R_{s \downarrow } }}$$
(1)

Net radiation was computed from the balance between short-wave and long-wave energy, by using the following formula:

$$R_{n} = R_{s \downarrow } - R_{s \uparrow } + R_{L \downarrow } - R_{L \uparrow }$$
(2)

where long-wave radiation emitted by the surface (RL↑, W m−2) was computed using the following equation (Brutsaert 1982):

$$R_{L \uparrow } = \varepsilon \sigma (T_{c} + 273.16)^{4}$$
(3)

where ε is the emissivity, σ is Stefan–Boltzmann constant (W m−2 K−4) and T c is the air temperature at canopy height (oC). Emissivity was set 0.98 for all stations. For emissivity values, see Gates (1980).

Net radiation was partitioned into latent (LE), sensible (H) and ground (G) heat fluxes (Penman 1948):

$$R_{n} = LE + H + G$$
(4)

The Bowen ratio, used to estimate the balance between sensible and latent heat fluxes, was expressed in the form (Bowen 1926):

$$\beta = \frac{H}{LE} = \gamma \frac{{T_{c} - T_{a} }}{{e_{c} - e_{a} }},$$
(5)

by approximating the fluxes by means of the temperature and humidity gradient. In the equation above, γ is the psychrometric constant (kPa K−1), T c −T a is the temperature difference between the air temperature at 2 m above the surface and the air temperature at canopy height (0.3 m), e c −e a − is the difference between the water vapour pressures (kPa) at these levels.

The water vapour pressure e (kPa) was computed by the formula:

$$e = \frac{{RH \cdot e_{w} }}{100}$$
(6)

where RH is the relative air humidity, e w is the saturation pressure of saturated water (kPa) in the air and at the canopy level, respectively.

The values of saturation pressure were obtained using the modified empirical Magnus–Teten’s equation (Buck 1996):

$$e_{w} = 0.61121exp\left( {\left( {18.678 - \frac{T}{234.5}} \right)\left( {\frac{T}{257.14 + T}} \right)} \right)$$
(7)

where T represent the temperatures at 2 m above the surface and at canopy level: (Ta) and (Tc), respectively.

The water vapour pressure deficit (VPD, kPa) was computed for 2 m above the surface level by the formula:

$$VPD = e_{w} - e_{a}$$
(8)

The ground heat flux was obtained applying Fourier’s law of heat conduction using the vertical method (Oke 1987; Monteith and Unsworth 1990):

$$G = k\frac{{T_{s} - T_{0.2} }}{{z_{s} - z_{0.2} }}$$
(9)

where k is the thermal conductivity of soil (W m−1 K−1); Ts and T0.2 are the soil temperatures at depths z s and z 0.2, respectively.

Thermal conductivity depends upon soil moisture content, mineral composition, dry soil density, particle composition, particle size distribution and soil temperature (Wierenga et al. 1969). Experimental thermal conductivity values were published by De Vries (1963); Kimball et al. (1976a, b; Asrar and Kanemasu (1983); De Vries and Philip (1986); Gregory et al. (1991); Sikora and Kossowski (1993); Peters-Lidard et al. (1998); and Ochsner et al. (2001).

The ground heat flux at the concrete surface was computed using an empirical equation due to the difficulty of positioning the thermometers within the concrete and isolating them from contact with the air (Brutsaert 1982).

$$G = cR_{n}$$
(10)

where R n is the net radiation and c is a constant (for bare soil = 0.3).

The latent heat flux was calculated by Bowen (1926):

$$LE = \frac{{R_{n} - G}}{1 + \beta }$$
(11)

Then, the sensible heat flux was computed on the basis of difference using the energy balance equation:

$$H = R_{n} = G - LE$$
(12)

Fluxes were considered positive when directed downwards and negative when directed upwards in relation to canopy height.

The evapotranspiration rate (ET, g m−2 s−1) was computed on the basis of the latent heat flux in the following way:

$$ET = \frac{LE}{{L_{e} }}$$
(13)

where L e is the latent heat of evaporation (J g−1). Daily sums of ET were expressed in mm.

The ratio of the incoming energy fluxes to the net radiation was used for the comparative analysis of the study sites. Furthermore, the evaporative fraction (EF, rel.) was used for assessing the amount of available energy consumed by evaporation (Lhomme and Elguero 1999; Suleiman and Crago 2004; Gentine et al. 2007). EF was computed using the equation:

$$EF = \frac{LE}{{R_{n} - G}}$$
(14)

A decoupling coefficient (Ω, unitless) was used for assessing the coupling between vegetation and the atmosphere. It was computed using the following equation (Jarvis and McNaughton 1985):

$$\varOmega = \frac{LE}{{LE_{p} }} = \frac{\varDelta + \gamma }{{\gamma + 1\left( {\frac{{r_{c} }}{{r_{a} }}} \right)}}$$
(15)

where LE p is the flux of potential evaporation (W m−2), Δ is the ratio between the saturation water vapour pressure gradient and the temperature gradient (kPa.°C−1), r c is bulk surface resistance (s m−1) and r a is aerodynamic resistance (s m−1). According to Jarvis and McNaughton (1985), the decoupling coefficient (factor) describes how closely the saturation deficit at the canopy (or leaf) surface is linked to that of the air outside the canopy boundary layer. It describes the sensitivity of evaporation to stomatal or surface conductance (Jones 1992). The decoupling coefficient is a dimensionless factor that assumes a value comprised between 0 and 1.

Aerodynamic resistance was computed using the Thom equation (Thom 1975):

$$r_{a} = \frac{{ \left[ ln\left( {\frac{z - d}{{z_{0m} }}} \right) - \psi_{m} (\varsigma )\right] \left[ {ln\left( {\frac{z - d}{{z_{0h} }}} \right) - \psi_{h} (\varsigma )} \right]}}{{Uk^{2} }}$$
(16)

where z is the height measurement of which was taken to describe the physical conditions of the air (m), d is the displacement height (m), z 0m and z 0h are aerodynamic roughness parameters referring, respectively, to the momentum and to heat transfer (m), Ψ m (ς) and Ψ h (ς) are the stability coefficients referring, respectively, to the momentum and to heat transfer (unitless), ς is the Monin–Obukhov stability parameter, U is the wind speed (m s−1) and κ is the von-Kármán constant (unitless).

The displacement height was computed as given below (Allen et al. 1998):

$$d = \frac{2}{3}h$$
(17)

where h is vegetation height. z 0m and z 0h were computed according to Allen et al. (1998):

$$z_{0m} = 0.123h$$
(18)
$$z_{0h} = 0.1z_{0m}$$
(19)

Stability parameters Ψ m (ς) and Ψ h (ς) were computed for stable (ς ≥ 0) and unstable or near neutral atmospheric conditions (ς < 0) (Foken 2008). Stability parameters for unstable and near neutral atmospheric conditions were computed in the following way (Liu et al. 2007):

$$\varPsi_{m} \left( \varsigma \right) = 2ln\left( {\frac{1 + x}{2}} \right) + ln\left( {\frac{{1 + x^{2} }}{2}} \right) - 2\arctan \left( x \right) + \frac{\pi }{2}$$
(20)
$$\varPsi_{h} \left( \varsigma \right) = 2\ln \left( {\frac{{1 + x^{2} }}{2}} \right)$$
(21)

where

$$x = (1 - 16\varsigma )^{0.25}$$
(22)

The stability parameters for stable atmospheric conditions (where ς ≥ 0) were computed according to Beljaars and Holstag (1991):

$$\varPsi_{m} \left( \varsigma \right) = - \left[ {a\varsigma + b\left( {\varsigma - \frac{c}{d}} \right)\exp \left( { - d\varsigma } \right) + \frac{bc}{d}} \right]$$
(23)
$$\varPsi_{h} \left( \varsigma \right) = - \left[ {\left( {1 + \frac{2a}{3}\varsigma } \right)^{1.5} + b\left( {\varsigma - \frac{c}{d}} \right)\exp \left( { - d\varsigma } \right) + \left( {\frac{bc}{d} - 1} \right)} \right]$$
(24)

where a = 1, b = 0.667, c = 5, d = 0.35.

The Monin–Obukhov stability parameter was computed as follows:

$$\varsigma = \frac{z}{L}$$
(25)

where L is the Monin–Obukhov length (m) computed by means of the following equation (Kalma 1989):

$$L = \frac{{u_{*}^{3} pc_{p} (T_{a} + 273.16)}}{kgH} = \frac{{u_{*}^{2} (T_{a} + 273.16)}}{{kgT_{*} }}$$
(26)

where u * is the wind friction velocity (m s−1), ρ is the air density (kg m−3), cp is the specific heat at constant pressure (J kg−1 K−1), g is the acceleration due to gravity (m s−2) and T * is a scaling parameter in the boundary layer, analogous to the friction velocity (°C). The wind friction velocity was computed by the equation given below (Kalma 1989):

$$u_{*} = \frac{kU}{{ln\frac{z - d}{{z_{0m} }} - \varPsi_{m} (\varsigma )}}$$
(27)

and T * using the following formula (Kalma 1989):

$$T_{*} = \frac{{k(T_{a} - T_{s} )}}{{ln\frac{z - d}{{z_{0h} }} - \varPsi_{h} (\varsigma )}}$$
(28)

All stability parameters, the Monin–Obukhov length, the friction velocity and the scaling parameter T * were computed by means of the iterative procedure proposed by Itier [1980, cited in Kalma (1989)]. For more details, see Kalma (1989) and Liu et al. (2007).

Bulk surface resistance was calculated according to the Penman–Monteith equation (Jackson et al. 1981; Wallace 1995); we used the following modification:

$$r_{c} = r_{a} \frac{{\frac{{\gamma r_{a} \left( {R_{n} - G} \right)}}{{pc_{p} }} - \left( {T_{c} - T_{a} } \right)\left( {\varDelta + \gamma } \right) - VPD}}{{\gamma \left[ {\left( {T_{c} - T_{a} } \right) - \frac{{r_{a} \left( {R_{n} - G} \right)}}{{\left( {pc_{p} } \right)}}} \right]}}$$
(29)

The mean values reported in Tables 3, 4, 5, 6 and 7 were derived from 24-hour data. Linear analysis was used to assess existing correlation between time and individual meteorological parameters and between time and energy fluxes. A confidence interval of 95 % was used, and p values <0.05 were considered significant.

Table 3 Mean monthly incoming solar radiation (Rs↓, W m−2), reflected solar radiation (Rs↑, W m−2), net radiation (Rn, W m−2), albedo (α,  %) and Bowen ratio (β) during the study period
Full size table

Results

Micrometeorological parameters

The monthly values of incoming solar radiation (Rs↓) for year 2008 (from January to December) were 145, 146, 132 and 123 kWh m−2 for May, June, July and August, respectively (Fig. 2).

Fig. 2
figure 2

Monthly values of incoming solar radiation (kWh m−2) data recorded for the January–December 2008 period. Mean values measured at the four sites

Full size image

The daily time courses of Rs↓ were similar at all sites and reached a maximum of about 900 W m−2 (Fig. 3a). The monthly average fluxes of Rs↓ (W m−2) for all the sites (Table 3) were the highest in June when mean values fluctuated from 300 W m−2 in the wet meadow to 314 W m−2 in the pasture. The average values of Rs↓ on sunny days for time period from May to August ranged between 291 (wet meadow) and 300 W m−2 (pasture).

Fig. 3
figure 3

Daily mean time series of incoming solar radiation (W m−2) (a) and reflected shortwave solar radiation (W m−2) (b), measured on clear days during the period: 1 May– 31 August, 2008

Full size image

The amount of reflected shortwave solar radiation (Rs↑) differed markedly between the different sites (Fig. 3b). The highest reflection, up to 230 W m−2, was measured over the concrete surface. The daily mean fluxes of Rs↑ at the remaining sites were quite similar to each other and ranged from 152 (pasture) to 175 W m−2 (wet meadow).

The monthly mean values of Rs↑ (Table 3) showed the highest mean flux over the concrete surface (81.7 W m−2). The mean fluxes of Rs↑ at the remaining sites were close together (from 59.4 to 62.5 W m−2). The diurnal mean time course of albedo, which is defined as the ratio of reflected to incoming solar radiation, (Fig. 4) showed 27 % at the site converted by concrete with a mean albedo for the sites with vegetation cover comprised between 20 and 21 %.

Fig. 4
figure 4

Daily series of albedo (%) measured on clear days at the four sites (the number of monitored days at each site is given in Table 2)

Full size image

The high albedo values measured in the early morning and in the late afternoon were both caused by a small angle of incident solar radiation and by dividing very small numbers of both variables of the ratio: they thus do not provide much useful information (Bray et al. 1966). We neglected these values and used for calculation the values between 6:00 and 18:00 h.

The diurnal course of net radiation (Rn) corrected for longwave fluxes (Fig. 5a) was quite similar at the three vegetated sites, not exceeding 600 W m−2. The maximum diurnal Rn on the concrete surface was lower, with a maximum of 430 W m−2.

Fig. 5
figure 5

Daily mean series of total net radiation (W m−2) (a), latent heat flux (W m−2) and evapotranspiration of water (mg m−2 s−1) (b), sensible heat flux (W m−2) (c), and ground heat flux (W m−2) (d) on clear days from 1 May to 31 August 2008 (the number of monitored days at each site are given in Table 2)

Full size image

The mean daily fluxes of Rn for all the sites (Table 3) ranged from 153 W m−2 (concrete surface) to 184 W m−2 (wet meadow and pasture). The lowest Rn was observed on the concrete surface, while the highest was over the vegetated sites.

The mean daily sums of Rs↓, Rs↑ and Rn calculated from the mean fluxes at four sites on clear days are given in Table 4. The mean daily values of Rs↓ ranged from 7.2 kWh m−2 at the pasture to 7.0 kWh m−2 at the wet meadow on clear days. The highest daily average of Rs↑ was 1.9 kWh m−2 at the concrete surface. The lowest average values of Rs↑ were observed at the barley field (1.4 kWh m−2). The average daily values of Rn fluctuated from 4.6 kWh m−2 (wet meadow) to 3.8 kWh m−2 (concrete).

Table 4 Mean daily incoming solar radiation (Rs↓), reflected solar radiation (Rs↑), net radiation (Rn), latent heat flux (LE), sensible heat flux (H) and ground heat flux (G) (kWh m−2)
Full size table

The daily courses of air temperature at canopy height (0.3 m) (Fig. 6a) showed that on clear days at 0.3 m, the pasture had the highest air temperature values; other sites showed similar midday temperatures, whereas the morning temperature differences between the localities were up to 8 °C. Low early morning temperatures at the wet meadow can be explained by the location of the wet meadow in a terrain depression. During the night, cold air originating from the surrounding higher ground, several metres above, flows down towards the wet meadow in the lowland surroundings. The early morning fog keeps temperature low and also provides more moisture.

Fig. 6
figure 6

Daily mean time series of air temperature (oC) (a) and relative air humidity (%) (b), measured at canopy height (0.3 m) on clear days calculated from 1 May to 31 August 2008

Full size image

On clear days, the average difference of air temperature measured at 0.3 m between the wet meadow and the concrete surface was 3.6 °C (Table 5). The average air temperatures’ differences at 2-m height between the three vegetated sites was 0.6 °C (17.8 °C at the wet meadow and 18.4 °C at the pasture) (Table 5). The maximum deviation at 2-m height between the vegetated sites and the concrete surface reached 1.8 °C (17.8 °C at the wet meadow and 19.6 °C at the concrete surface). The air temperature at 0.3 m at the barley field varied from 15.1 °C (May) to 20.8 °C (August), while the canopy air temperature at the wet meadow ranged from 14.6 to 17.2 °C.

Table 5 Monthly mean air temperature (oC ± SD) and relative air humidity (% ± SD) at 2 m (Ta, RHa) and 0.3 m (Ts, RHs) for all sites from 1 May to 31 August 2008
Full size table

The time course of the average values of relative air humidity (RH) at 0.3 m (Fig. 6b) shows the highest values at the pasture (about 80 %) and a similar pattern at the wet meadow. The mean values of RH at 2 and 0.3 m (Table 5) showed the highest RH (80 %) at 0.3 m in the pasture and the lowest (75 %) in the field. In contrast, the pasture, the wet meadow and the field had values of mean RH close to each other at 2 m on clear days (67.7, 69.0 and 68.4 %, respectively). Differences in RH were probably caused by a different development of the vegetation cover (above ground biomass, LAI) and of the soil water content during the vegetated season.

Variation in energy fluxes

The way in which Rn was processed differed amongst sites (Table 6). The energy involved in evapotranspiration ranged between 25.6 W m−2 (concrete surface) and 164.7 W m−2 (wet meadow). The monthly mean latent heat fluxes (LE) values were similar in the field and in the pasture (134.8 and 143.5 W m−2, respectively). The ground heat fluxes (G), which represent the smallest component of Rn, ranged from 7.1 W m−2 (pasture) to 38.5 W m−2 (concrete surface). The evaporative fraction (EF) [i.e. the ratios of evapotranspiration to available energy at the ground surface (Rn − G)] ranged from 0.22 (concrete) to 1.04 (wet meadow).

Table 6 Monthly mean values of energy balance component, evaporative fraction (EF) and evapotranspiration (ET) during the measurement period
Full size table

The hourly energy fluxes measured on clear days at canopy height level in the four sites (Fig. 5) reached value of several 100 W m−2. At all sites, LE flux peaked in the afternoon, while the minimum was observed during the night. Mean daytime LE reached a maximum of about 630 W m−2 (wet meadow). Mean diurnal LE reached similar values (ca. 400 W m−2) at the pasture and at the field, while at the concrete surface it did not reach 230 W m−2. At the wet meadow, most of Rn was converted into LE; conversely, at the concrete surface, sensible heat (H) was the dominant flux measured (up to 370 W m−2). At all vegetated sites, H was relatively low (less than 220 W m−2), whereas LE fluxes reached a maximum comprised between 400 and 600 W m−2, which corresponds to ET rates of up to 0.15 g m−2 s−1. The highest diurnal G was up to 150 W m−2 at the concrete surface, 40 W m−2 at the wet meadow, 40 W m−2 at the field and 18 W m−2 at the pasture.

The mean daily sums of LE, H and G calculated from the mean fluxes measured at the four sites on clear days are reported in Table 4. The mean daily LE values ranged from 2.7 kWh m−2 at the wet meadow to 0.5 kWh m−2 at the concrete surface on clear days. The highest daily average H was 2.9 kWh m−2 at the concrete surface, while the lowest average H was observed at the wet meadow (1.6 kWh m−2). The daily average G values ranged from 0.2 kWh m−2 (pasture) to 1.1 kWh m−2 (concrete).

The dissipation of Rn was calculated for each site (Fig. 7). The wetland had the highest LE/Rn (0.90), while midrange values of LE/Rn were recorded at the pasture and at the field (0.79 and 0.75, respectively), while the artificial concrete surface had the lowest LE/Rn (0.26). At all sites, G/Rn ranged from 0.06 to 0.10.

Fig. 7
figure 7

Dissipation of net radiation (%) amongst latent heat, sensible heat and ground heat fluxes on clear days from 1 May to 31 August 2008

Full size image

The mean aerodynamic resistance (ra), bulk surface resistance (rc), vapour pressure deficit (VPD), available energy (Rn − G) and decoupling coefficient (Ω), assessed during three separate periods (Table 7), were in the ranges of 122–257, 73–229, and 96–315 s m−1 for the field, the wet meadow and the pasture, respectively, between 10:00 and 16:00. The relationship amongst surface resistance, available energy, vapour pressure deficit and air temperature at the three vegetated sites (Fig. 8) showed that rc increased linearly with the increasing VPD (R 2 = 0.60, df = 1, 289, p < 0.001 (field); R 2 = 0.46, df = 1, 301, p < 0.001 (wet meadow); R 2 = 0.55, df = 1, 291, p < 0.001 (pasture) (Fig. 8a). The surface resistance also increased linearly with the increasing air temperature (R 2 = 0.32, df = 1, 289, p < 0.001 (field); R 2 = 0.44, df = 1, 301, p < 0.001 (wet meadow); R 2 = 0.41, df = 1, 291, p < 0.001 (pasture) (Fig. 8b). Surface resistance rose in parallel to an increase in available energy (Fig. 8c); however, the slope of the regression line was low.

Table 7 Diurnal average values of wind velocity (uz), surface (rc, s m−1) and aerodynamic resistance (ra, s m−1), vapour pressure deficit (VPD, kPa), available energy (Rn − G, W m−2) and decoupling coefficient (Ω) for three periods between 10:00–16:00
Full size table
Fig. 8
figure 8

Relationship between surface resistance (s m−1) and water VPD (kPa) (a), air temperature (oC) (b), and available energy (W m−2) (c) for the three vegetated sites

Full size image

Mean ra values had a small variation and were in the ranges of 30–61, 35–48, and 41–80 s m−1 at the field, the wet meadow and the pasture, respectively. The mean values of wind speed did not exceed 3.5 m s−1. The decoupling coefficient varied little between different periods (Table 7). Mean values were between 0.60 and 0.83, 0.65 and 1.23, 0.80 and 0.89 for field, wet meadow and pasture, respectively. The highest mean values of Ω were observed in July at the field and at the pasture. At the wet meadow, in August, Ω exceeded 1.

Discussion

Monitored data and evaluated energy fluxes

Evapotranspiration occupies a central role in the processing of water and energy fluxes within ecosystems. The most powerful fluxes of energy dissipation on clear days are linked to sensible heat and/or to latent heat of evapotranspiration. Our results clearly show how, depending on the availability of water and vegetation, different land cover types convert solar energy in different ways. During dry spells, at the wet meadow, more than 100 % of net radiation is dissipated through evapotranspiration. These values indicate that at the wet meadow, up to 70 % more energy was converted into latent heat of evapotranspiration than at the concrete surface site, and up to 30 % more than at the pasture or at the barley field. Many physical and biological factors, such as leaf area index, available energy, soil moisture ability, plant physiology, production characteristics and vapour pressure deficit, bear an influence on energy dissipation (Liu et al. 2009). A high dissipation of latent heat flux (LE/Rn) at the wet meadow site can be explained because of high moisture availability and to low surface resistance to evaporation. The midrange values recorded at the field and at the pasture sites are due to the nature of their vegetation cover and to the higher surface resistance which counterbalanced the high amount of available moisture. Small dissipation of LE/Rn values at the concrete surface site are caused by the absence of vegetation and the small water holding capacity. On the whole, at all vegetated sites, LE/Rn exceeded 0.70. These data demonstrated that evapotranspiration is a major component of the energy balance under dry hot days in our region. Similar results were found at the tropical prairie wetland in north-central Florida, USA, where approximately 64 % of net radiation was dissipated through evapotranspiration, and 32 % was converted into sensible heat (Jakobs et al. 2002).

Evapotranspiration rates which we measured at the wet meadow ranged from 3.3 to 7.1 mm d−1. In similar geographical zone, Burba et al. (1999) measured 3.5–6.5 mm d−1 at a prairie wetland site in Nebraska, USA. Peacock and Hess (2004) measured 0.5–5.5 mm d−1 in Kent, UK, while Zhou and Zhou (2009) observed between 0.1 and 5.8 mm d−1 in reed beds during the growing season in the Liaohe Delta, Northeast China. By comparison, evapotranspiration rates over the Amazonian forest varied between 2 and 6 mm d−1 and between 0.4 and 5 mm d−1 during the wet and the dry seasons, respectively (Sanches et al. 2011).

Blad and Rosenberg (1976); Brakke et al. (1978); Monteith (1981); Oke (1987); Guo and Schuepp (1994); Lee et al. (2004); Ryszkowski and Kedziora (2007) showed that local heat advection contributed to increased evaporation and/or evapotranspiration. Advection takes place when sensible heat flux is negative at or near the ground level (McNaughton and Jarvis 1983). Horizontal advection transfers heat from higher to lower temperatures because of the canopy–air temperature differences (Li and Yu 2007). Our results showed such a temperature gradient between the agricultural landscapes and the wet meadow. The negative sensible heat values observed at the wet meadow site indicate the presence of temperature inversion caused by the advection of warm air from surrounding areas which tends to occur during clear days. Guo and Shuepp (1994) stressed that the evapotranspiration at wetland sites can be enhanced by local heat advection from neighbouring areas by more than 20 %. Brakke and Verma (1978) found that sensible heat advection provided about 50 % of the total energy dissipated through evapotranspiration. The process of dry air advection from a field to a hedgerow was described by Ryszkowski and Kedziora (2007). At the cultivated field, more net radiation was converted into sensible heat than at the forest. A temperature gradient was measured between the field and the hedgerows proving that heat advection occurred. Our results correspond well to the results obtained at a temperate wetland under hot summer conditions measured by Rejšková et al. (2010). In their study, the evaporative fraction reached values higher than 1.0 in the afternoon hours of a hot day because of advection.

The surface reflectance (albedo) depends on the vegetation cover, namely, on its biomass and on its water content. Lafleur et al. (1987) reported an average albedo of 11–20 % for sedge-dominated wetlands, while Burba et al. (1999) reported an albedo of 12–16 % for common reed. In our study, the albedo for the vegetated sites is in agreement with the results by Brunsell et al. (2011), where the variation of the albedo ranged from 18 to 25 % as a function of vegetation cover.

Low albedo was associated with high values of LE/Rn (wet meadow), while high albedo was associated with low LE/Rn. The values of EF varied from site to site; at the wet meadow, EF was the highest with 1.04, a value which is well above the EF (0.84) reported at the wet meadow studied by Ryszkowski and Kedziora (1987). At the concrete surface, EF was 0.30, which was comparable to the urban surface (0.29) reported by Coutts et al. (2007). At the pasture, EF was slightly higher (0.82) in comparison to the EF values of observed at the pasture (0.55) by Wang et al. (2006). This difference could be explained by a higher groundwater level at our pasture site.

Crop resistance (rc) is related to the available energy for evapotranspiration (Rn − G), temperature gradient, the vapour pressure deficit (VPD) and aerodynamic resistance of the surface (Jackson et al. 1981; Wallace 1995). The decrease of rc is related to an increase in irradiance, and it rises with increasing VPD (McNaughton and Jarvis 1983). Our results indicate that rc is more sensitive to VPD and to temperature than to the amount of available energy. Doorenbos and Pruitt (1977) reported a standard rc value for grass of 70 s m−1. The rc above the wet grassland fluctuated between 50 and 100 s m−1 (Acreman et al. 2003; Jakobs et al. 2002). Wessel and Rouse (1994) reported that rc can vary between as much as 29 and 251 s m−1.

Effect of ET on local climate and water cycle

Our results document a striking difference in the seasonal energy balance measured at the four sites with only a few kilometres between each other. Under hot weather conditions, the wet meadow showed a high capacity of energy conversion into latent heat of water vaporisation. Wetlands can compensate for temperature rises because of warm dry air by increasing their evapotranspiration rate, which can be even higher than the value of net radiation at the wetland site. Different ET rates result in differences in land covers surface temperature at vegetation (canopy) height. In other words, the temperature measured at the height of the dominant vegetation in various vegetated sites can serve as an indicator of the rate of evapotranspiration. A detailed study of daily dynamics of radiation surface temperature of seven types of land cover in a cultural temperate landscape showed a difference of up to 20 °C between different types of vegetation supplied with water and drained areas (Hesslerová et al. 2013).

The excess of heat and the lack of water over arable fields dramatically reduce crop yields, while wetlands play a very important role in moderating the water cycle and retaining water within the catchment. The structure of agricultural landscapes bears important linkages with the local climate. Restoring wetlands next to agricultural plots could be an effective measure for managing the heat balance of the landscape. The average maximum ET during clear days is a demonstration of the role of vegetation in dampening temperature differences caused by surplus solar radiation and in shaping microclimate and mesoclimate conditions. Within wet meadows, ET reaches ca. 600 W m−2, which when projected over 1 km−2, corresponds to 600 MW—the efficiency of a midsized electricity generating power station. ET equalises temperature differences within the landscape both in space and in time. Transpiration is driven by conversion of solar energy into latent heat of water vaporisation. Energy used for water vaporisation does not increase ambient temperature; the energy is hidden (latent) in kinetic movement of molecules of water vapour. When temperature drops below the dew point, water vapour condensates, and latent heat is released. Transpiration is a cooling process (endothermic), whereas condensation of water vapour is on the opposite a warming process (exothermic). The water cycle (water vaporisation–water vapour condensation) is a process equalising temperature differences (lowering heat potentials); plants can be considered as processors controlling the equalisation of temperature differences. There are up to several 100 stomata for every mm2 of leaf surface; each stoma functions like a valve controlling the uptake of carbon dioxide and the release of water vapour.

The cooling process of transpiration is often considered a side effect rather than a mechanism to control leaf temperature (Lambers et al. 1998). Transpiration is also perceived as a rather negative process. Plant physiologists and hydrologists may use negative terms such as ‘transpiration loss’ and ‘evapotranspiration losses’.

A similarly great controversy exists at present in the discussions about the functioning of forests’ and their role in the hydrological cycle. Transpiration is sometimes even considered an unavoidable evil, in the sense that water is sacrificed for the sake of enabling intake of CO2 for photosynthesis to occur. When comparing relatively small catchments, less rainfall is converted into runoff from afforested catchments than from grass covered or partly drained catchments (Andreassian 2004). On the other hand, large scale deforestation has often been linked to less precipitation and to regional shortages of water (Ponting 1993, Diamond 2005).

Makarieva and Gorshkov (2007) point out that, in forest-covered regions, annual precipitation does not decline with the increasing distance from the ocean, and may even rise as one proceeds several 1000 km inland. By contrast, where forests are lacking, precipitation decreases exponentially over just a few 100 km. They coined the term ‘biotic pump’ of atmospheric moisture to designate the biotically induced atmospheric circulation sustaining the hydrological cycle on land (Makarieva and Gorshkov 2007). Their initial theory has more recently been supported by new evidence (Makarieva et al. 2013). In this last publication, they point to the inversion of temperature that occurs in a structured forest canopy in comparison with simple crop plants: there is higher temperature in tree crowns than in the herb and shrub layers. As heavier, colder air remains at ground level, water vapour may condense over herbs and shrubs even during a sunny day. When the temperature goes down at night, the air becomes more saturated and condensation occurs above the tree canopy. Condensation of water vapour results in a decrease of air pressure, such that air from surrounding areas is sucked in. According to Makarieva and Gorshkov (2007), vegetate dares as characterised by intense evapotranspiration constitute ‘acceptor regions’ of low atmospheric pressure on land which attract water vapour. On the other hand, dry, overheated areas function as ‘donor regions’; they lose water transported in turbulent fluxes of hot air (sensible heat). Such overheated areas are prone to the rapid decomposition of organic material, loss of nutrients and erosion (Ripl 2003; Hesslerová et al. 2013).

In an agricultural landscape, wetlands provide a climate regulation service, which is linked to closed water cycles, nutrient retention, biomass and oxygen production (Seják et al. 2012). Climate regulation is a critical regulatory ecosystem service provided by wetlands, which dampens temperature differences within the landscape. This ecosystem service can be expressed in monetary terms by relating it to the cost of the electricity that would be consumed for providing this air-conditioning effect artificially. Each m2 from 500 l water evaporate annually would require an investment of 70 USD (i.e. 0.7 kWh cooling + 0.7 kWh warming × 0.1 USD × 500 l).

The Unites States have lost over half of their wetlands since Independence was declared (Dahl 2011). The wetland surface area deceased from 200 million ha in 1780 to 100 ha in 2,000—a loss of 247 million acres. Seven states—Indiana, Illinois, Iowa, Missouri, Ohio and California—have lost over 90 % (http://www.wetland.org/101/WET101C.pdf). Johnston (2013) found wetland losses continued at a rate of 15,000 acres a year between 2001 and 2011 because of row crop expansion. In the Czech Republic, about 1 million ha of wetlands and wet meadows were drained between the 1950s and the end of the 1980s under the Socialist Regime. The decrease of ET of 100 W m2 corresponds to release of 100 MW of sensible heat from 1 km2. The drainage of a million ha (10,000 km2) results in the release of 1 million MW of sensible heat on sunny days, which increases temperature and accelerates the transport of water vapour into the higher layers of the atmosphere. The restoration of wetlands in the agricultural landscape is therefore necessary not only for the retention of nutrients (N, P) and for carbon sequestration, but is also essential for the dampening of heat extremes and for equalising temperature differences.