Modeling Long-Term Temporal Variation of Dew Formation in Jordan and Its Link to Climate Change

: In this study, we performed model simulations to investigate the spatial, seasonal, and annual dew yield during 40 years (1979–2018) at ten locations reﬂecting the variation of climate and environmental conditions in Jordan. In accordance with the climate zones in Jordan, the dew formation had distinguished characteristics features with respect to the yield, seasonal variation, and spatial variation. The highest water dew yield (an overall annual mean cumulative dew yield as high as 88 mm) was obtained for the Mountains Heights Plateau , which has a Mediterranean climate. The least dew yield (as low as 19 mm) was obtained in Badia , which has an arid climate. The dew yield had a decreasing trend in the past 40 years due to climate change impacts such as increased desertiﬁcation and the potential of sand and dust storms in the region. In addition, increased anthropogenic air pollution slows down the conversion of vapor to liquid phase change, which also impacts the potential of dew formation. The dew yield showed three distinguished seasonal patterns reﬂecting the three climates in Jordan. The Mountains Heights Plateau (Mediterranean climate) has the highest potential for dew harvesting (especially during the summer) than Badia (semi-arid climate).


Introduction
Climate change impacts on water resources and supplies. Disruption of water supplies has socio-economic consequences that might take many years to recover. By the year 2025 more than 50% of the world population will suffer from freshwater supplies [1]. Therefore, awareness, international agreements, and national strategies are built up worldwide to wisely manage and restore water resources on both local and regional scales [2]. Technological advancement also plays an important role in decreasing the effects of climate change on water resources. For example, water desalination has been utilized whenever it is affordable to produce fresh water. Alternatively, cloud seeding by injecting cloud condensation nuclei (CCN) or ice nuclei (IN) in the atmosphere may also help in improving rain patterns in some semi-arid and arid areas. Recently, dew and fog water harvesting has been introduced in semi-arid areas as a source of water [3][4][5][6][7][8][9][10][11].  Table 1 for locations characteristics.   Table 1 for locations characteristics. Jordan is a small country (~89,000 km 2 , population~10 million in 2016). It is located (29 • -34 • North and 34 • -40 • East) at the Asian-African continental border-line and about 100 km east of the Mediterranean Sea. It has diverse habitats, ecosystems, biota due to its varied landscapes, environments, and climates.
The summer season in Jordan is hot-dry and spans from May to September with a mean temperature of~32 • C (sometimes exceeds 40 • C in July and August). The winter is relatively cool and Water 2020, 12, 2186 4 of 18 spans from November to March with mean temperature~13 • C and frequent showers and occasional snowfall in some elevated areas.
Geographically, Jordan comprises a wide variety of topography that defines its climate spatial variation [52,53]. There are three distinguished climatic zones in Jordan: (1) Jordan Valley, (3) the Eastern Desert (also known as Badia), and (3) Mountains Heights Plateau. The Jordan Valley (including Jordan River and the Dead Sea) climate ranges from arid to semi-arid with a hot dry summer, warm winter, and precipitation less than 200 mm/yr. The climate in the Badia is characterized by a sharp change in temperature between day and night and between summer and winter. For example, the daytime summer temperature in Badia can exceed 40 • C and the winter nights can be very cold, dry, and windy. Precipitation in Badia is less than 100 mm/year.
The Mountains Heights Plateau (including highlands above the Jordan Valley, mountains of the Dead Sea, Wadi Araba, and Ras Al-Naqab) are mainly situated in the western region starting at the northern part parallel to the Jordan Valley and extended to the south approaching Wadi Rum and Aqaba. The climate on the Mountains Heights Plateau is a typical Mediterranean climate with a hot-dry summer and cool-wet winter with two short transitional seasons. The Mountains Heights Plateau receives Jordan's highest amounts of precipitation (more than 300 mm/year), which falls during October-May with the peak usually during winter (December-February).

Fresh Water Resources
Jordan is the fourth country in the world suffering from freshwater shortages. The main sources of water in Jordan include safe abstraction of groundwater, recycling wastewater, surface runoff water, and desalination. The annual mean water amount received in the form of rainfall is about 8300 million cubic meters (MCM) [54]. The available water per capita has declined considerably during the past century; it was about 3600 m 3 in 1946 and it is expected to be as low as 100 m 3 in 2025 [55]. Jordan's water demand estimated to be about 940 MCM (63% agriculture, 32% domestic, and 5% industry) in 2007 and it increased to be about 1600 MCM in 2010. Groundwater capacity consists of twelve identified basins with the critical state due to maximum capacity exploitation. The Jordanian authorities have built ten water dams to utilize the stored water and satisfy livestock needs and groundwater recharge [55]. In addition, surface water has a share of about 28% of the total water supply in the country.

Spatial Variation of the Cumulative Dew Yield
We calculated the overall average cumulative yield of dew in liquid form (i.e., condensation as liquid water) and solid form ( Table 2). The conditions to calculate dew in the solid phase (i.e., hoarfrost) were described in detail in the methods section. In brief, this happens when T c < 0 • C. This condition was rarely valid almost for all sites (i.e., less than 1 mm). The only exception was in Mafraq, where the mean hoarfrost yield was~2 ± 1 mm.
The highest overall average water dew yield was obtained for the site on the Mountains Heights Plateau, which has a Mediterranean climate. At these sites, the overall average water dew yield was 46 ± 6, 40 ± 6, 55 ± 6, 57 ± 6, and 88 ± 7 mm, respectively, for Petra (site 8), Karak (Site 7), Amman (Site 3), Madaba (Site 4), and Irbid (Site 1). These regions have a Mediterranean climate with humid winter and hot dry summer. Average air temperature, dew point temperature, and relative humidity in these sites are 19 • C, 9 • C, and 58%; respectively ( Table 1). The lowest overall average water dew yield was obtained for the sites located in the Badia had the least overall average water dew yield: Ma'an (site 9, 19 ± 3 mm), Mafraq (site 2, 19 ± 4 mm), and Azraq (site 5, 22 ± 4 mm). All these sites were located in the Eastern Desert climate (also known as Badia). The Dead Sea (site 6) also was one of the sites that showed a high value for the overall average water dew yield (about 63 ± 7 mm). Aqaba (site 10) had an overall average water dew yield of about 28 ± 3 mm. From a climatological point of view, the Aqaba site is affected by the desert climate. Table 2. Overall mean ± standard deviation as well as the range of the yearly cumulative dew yield. The values are listed as mean ± Std (min-max) in mm.

Site Number
Site Name Water Hoarfrost Water + Hoarfrost

Temporal Variation of the Cumulative Dew Yield
During the past two decades, dew yield has declined at all sites, especially in the arid and semi-arid sites (Figures 2 and 3). It is not surprising to see the model predicted such a decline during that period because the climate change impacts were very pronounced by increased desertification and potential of sand and dust storms (SDS) in the region [56,57] as will be discussed later.

Temporal Variation of the Cumulative Dew Yield
During the past two decades, dew yield has declined at all sites, especially in the arid and semi-arid sites (Figures 2 and 3). It is not surprising to see the model predicted such a decline during that period because the climate change impacts were very pronounced by increased desertification and potential of sand and dust storms (SDS) in the region [56,57] as will be discussed later.    The Mann-Kendal trend analysis [58] at confidence level 95% (p < 0.05) revealed a significant negative trend for the yearly mean dew yield at all selected sites (Table 3). To find out the reason for such a decreasing trend in dew yield, we investigated the trend analysis results for air temperature, dew point temperature, and relative humidity. These are key factors in dew formation. Air temperature (T a , T max, and T min ) and relative humidity showed a significant positive and negative trend during the 40 years (Tables 4 and 5). T d did not show any significant trend. Nevertheless, it seems that the negative trend in dew yield was a result of the increase in temperature and decrease in relative humidity in the past 40 years.
Ma'an −0.1 * −1 * Values with a star indicate a statistically significant trend with p < 0.05.
Ma'an −0.09 * −0.   Again, these patterns reflected the three distinguished climate zones in Jordan. For example, Pattern-I was seen for Mafraq (site 2), Azraq (site 5), Ma'an (site 9), and also Aqaba (site 10); these sites are all within the arid climate in the Badia zone. Pattern-II was mainly seen for Amman (site 3), Madaba (site 4), Karak (site 7), and Petra (site 8); all are within the Mediterranean climate zone on Mountains Heights Plateau. Although the Dead Sea (site 6) is located within the semi-arid climate zone (Jordan Valley), but it showed a similar seasonal pattern as that seen for the sites on the Mountains Heights Plateau. Even though Irbid (site 1) is a site with a Mediterranean climate as it is located on the Mountains Heights Plateau zone, but it had a very distinguished seasonal pattern (i.e., Pattern-III), Water 2020, 12, 2186 9 of 18 which had elevated values of dew yield during the summer and autumn. This is not surprising because Irbid (site 1) is also located at the edge of the Fertile Crescent, which receives the highest amounts of rainfall and it seems to be a site favoring dew formation as well.
Water 2020, 12, x FOR PEER REVIEW 9 of 19 Mountains Heights Plateau. Although the Dead Sea (site 6) is located within the semi-arid climate zone (Jordan Valley), but it showed a similar seasonal pattern as that seen for the sites on the Mountains Heights Plateau. Even though Irbid (site 1) is a site with a Mediterranean climate as it is located on the Mountains Heights Plateau zone, but it had a very distinguished seasonal pattern (i.e., Pattern-III), which had elevated values of dew yield during the summer and autumn. This is not surprising because Irbid (site 1) is also located at the edge of the Fertile Crescent, which receives the highest amounts of rainfall and it seems to be a site favoring dew formation as well.

Discussion
The difference in the dew formation yield among selected sites reflected the climate variation described before in the methods section; i.e., Jordan Valley zone (site 6), Eastern Desert (also known as Badia) zone (sites 2, 5, and 9), and Mountains Heights Plateau zone (sites 1, 3, 4, 7, and 8). Aqaba (site 10) was a coastal site, but it was not one of the sites with high overall average dew yield due to its location very close

Discussion
The difference in the dew formation yield among selected sites reflected the climate variation described before in the methods section; i.e., Jordan Valley zone (site 6), Eastern Desert (also known as Badia) zone (sites 2, 5, and 9), and Mountains Heights Plateau zone (sites 1, 3, 4, 7, and 8). Aqaba (site 10) was a coastal site, but it was not one of the sites with high overall average dew yield due to its location very close to the arid and desert region; the higher the temperature is the higher potential to keep the water vapor in the gas phase and that prevents condensation on surfaces (i.e., dew formation).
As for the dew yield decline during the past two decades (2005-2018) and the climate change impacts, two previous studies reported increased desertification and potential of sand and dust storms (SDS) in the Middle region due to warming and drying episodes [56,57]. For example, warming and drying episodes increased during that period [56]. In turn, these have significant impacts on the albedo and short-wave radiation leading to higher surface reflection [59], which directly affects the potential of dew formation. Besides climate change impacts, the increased emissions of anthropogenic air pollution impact the potential of dew formation [60] because submicron particles slow down the conversion of water vapor from the gas phase to the liquid phase. This, in turn, escalated the desertification process in the Middle East causing increased frequency of dust episodes and atmospheric dust particles concentrations, which affected the albedo and the short-wave radiation as per the discussion above. The processes involved in dew formation can be similar to that involved in the cloud droplet formation with a difference that clouds are formed on by condensation on supercooled particles (i.e., cloud condensation nuclei CCN) whereas dew is formed by condensation on environmental surfaces with their temperature lower than the dew point. Therefore, anthropogenic air pollution might also similarly hinder dew formation.
To have an insight into seasonal patterns of dew yield, we investigated the long-term mean of the seasonal variation for some meteorological parameters (T a , T d , RH, WS, surface pressure, and total cloud cover), which are key factors in dew formation (Figures S1-S10 in Supplementary Material). Regardless of the variation in the amount of dew yield, the peak of dew formation can occur in the wintertime (November-February). Indeed, during the cold season, temperature declines in all parts of Jordan (mean T a~1 0 • C). Besides that, the domination of the prevailing westerly winds brings humid air from the Mediterranean Sea into the country; consequently, the relative humidity rises, and the difference between the temperature and the dew point (T a -T d ) decreases. All these together with the increasing cloud cover, which is a result of the Mediterranean cyclones, provide the initial conditions for dew formation in the winter. Therefore, dew occurs frequently in most parts of Jordan during the wintertime.
Differences in the amount of dew yield are clear in different locations (~100-350 mL/m 2 ). Although the probability of dew occurrence in winter is mainly similar at all sites, but it is different in summer. In fact, the occurrence of dew in winter is mostly controlled by synoptic conditions whereas in summer it is more dependent on local conditions and is largely affected by the features of the climate zones. For instance, in Mafraq (site 2), Azraq (site 5), Ma'an (site 9), and Aqaba (site 10), which all were within the arid climate in the Badia zone, dew is almost vanished during late spring and summer (May-September). Indeed, in the summer, the temperature sometimes exceeds 40 • C (especially in July and August), and relative humidity is very low, the sky is often clear, the daytime duration is long, and the incoming short-wave radiation exceeds the outgoing long wave radiation. Furthermore, high surface temperature leads to form thermal low-pressure system, which cause turbulence and an intense wind speed. As such, condensation cannot occur, and dew is almost vanished during the warm season.
Dew occurred throughout the year in Amman (site 3), Madaba (site 4), Karak (site 7), and Petra (site 8) as a result of the moist air fronts prevailing from the Mediterranean Sea (i.e., sea breeze, especially after sunset). However, the Mediterranean Sea breeze does not reach beyond 100 km inland. Furthermore, these sites were located on high elevation; thus, the temperature drops in the night and T a -T d is reduced, which favors dew formation.
An interesting result was observed in Irbid (site 1). This site had a unique seasonal pattern of dew formation with a significant potential for dew formation throughout the year (~90 mm/year) and a maximum daily dew yield during summer and early autumn (July-September). This site has close surface water in the north (i.e., the Sea of Galilee) and receives the Mediterranean Sea breeze more efficiently than any other region in Jordan. The temperature is also moderate in this region during the summer, which another reason that favors dew formation in the summer. One reason for higher dew yield in summer than winter in this site could be due to air temperature. On the other hand, in wintertime due to lower temperatures, the atmosphere cannot keep a significant amount of moisture, whereas during warm season with a rise in temperature, the atmospheric capacity to keep water vapor increases. This means that high capacity of moisture can compensate for high temperatures in summer and lower T a -T d . All these conditions together with light wind speed (i.e., less than 2 m/s) stimulate dew formation in Irbid during the summer.

Conclusions
Water resources are very scarce in Jordan. There is a great need to look for additional water resources; and hence, dew can be one of such potential sources to be utilized for agricultural purposes. In this study, we performed model simulations to investigate the spatial, seasonal, and annual dew yield during 40 years (1979-2018). We considered ten locations selected in different geographical locations but have the same altitude to reflect the variation of climate and environmental conditions.
In accordance with the climate zones in Jordan, the dew yield had distinguished characteristics features with respect to the yield, seasonal variation, and spatial variation. The highest water dew yield was obtained for the Mountains Heights Plateau, which has a Mediterranean climate, with an overall annual average water dew yield was as high as 88 mm. The Badia, which has an arid climate, received the least dew yield with an overall annual average water dew yield was as low as 19 mm. During the past two decades, the dew yield was declined as a result of climate change impacts in the form of increased desertification and the potential of sand and dust storms (SDS) in the region. In addition, the increased anthropogenic air pollution impacted the potential of dew formation because submicron particles slow down the conversion of water vapor condensation on obstacles.
According to the model simulations, each climate zone in Jordan had a distinguished dew yield seasonal pattern. The first pattern was characterized by high yield value (daily cumulative exceeding 100 mL/m 2 ) during winter and almost vanishing yield during summer. The second pattern is characterized by significant amounts of daily cumulative dew yield (exceeding 100 mL/m 2 ) throughout the year with yields as high as 200 mL/m 2 during the winter. The third pattern was characterized by a rather constant daily cumulative dew yield (about 200 mL/m 2 ) during the winter and spring and high yield values (as high as 350 mL/m 2 ) during the summer and autumn.
The outcome of this research can be utilized for managing and planning local feasibility studies for dew harvesting in Jordan. Knowing the locations with the highest dew formation potential is important from a strategic point of view to make the suitable decision for commissioning dew water harvesting as a secondary source of water to be used in different sectors. In the future, model calibration is the most important part is validating the results in this study and carry it beyond selected locations and cover all parts of Jordan (i.e., gridded model simulation). Model validation requires long-term dew harvesting in different locations, which is laborious. The upcoming research in this context is to prepare for the long-term measurement campaign and redo the model simulation for a gridded domain covering whole Jordan. The global dew formation model, which was developed by Vuollekoski et al. [41], was modified to accommodate for environmental conditions in Jordan. The original model setup was made suitable for a global contest and can handle water phase change between vapor and liquid or solid (i.e., condensation and desublimation); only dew was considered but precipitation and fog were not. The substrate (i.e., condenser material) was assumed a horizontally aligned sheet (at 2 m height and thermally insulated from the ground) of a suitable material such as low-density polyethylene (LDPE) or polymethylmethacrylate (PMMA).
The model describes the water phase change based on mass and heat balance Equation where dT c /dt is the change rate in the condenser temperature. C c , C w , and C i are the specific heat capacity of condenser, water, and ice; respectively. Here, m c , m w , and m i are mass of condenser, water, and ice; respectively. The right-hand side describes the heat exchange involved in the heat exchange processes: P rad is the incoming and outgoing radiation, P cond is the conductive heat exchange between the condenser surface and the ground, P conv is the convective heat exchange, and P lat is the latent heat released by the condensation or desublimation of water. The model reads all input data for a given grid point and solves equations. using a fourth-order Runge-Kutta algorithm with a 10 s time step. All terms in Equation (A1) are described in more detail in Tables A1 and A2. The model was set up so that it assumes similar conditions for the phase-change of pre-existing water or ice on the condenser sheet. For instance, if the water on the condenser is in the liquid phase (i.e., m w > 0) and the condenser temperature T c < 0 • C, then the sheet is losing energy (i.e., the right-hand side of Equation (A1) is negative). In that case, instead of solving Equation (A1), Tc is assumed constant and the lost mass from the liquid phase of water is transferred to the cumulated mass of ice; i.e., the water is transformed from liquid phase to solid phase. Consequently, Equation (A1) is replaced by L wi dm w dt = P rad + P conv + P lat , where L wi [J kg −1 ] is the latent heat of fusion. If the water on the condenser is in the solid phase (i.e., m i > 0) and the condenser temperature T c > 0 • C, a similar equation is assumed for the change rate of ice mass (m i ).
Note that Equation (A2) is not related to the condensation of water; it only describes the phase change of the already condensed water or ice on the condenser. For the water condensation rate, which is assumed independent of Equation (A2), the mass-balance equation is then assumed to be where m represents either the mass of ice (m i ) or water (m w ) depending on weather Tc is below or above 0 • C. P sat (T d ) is the saturation pressure at the dew point temperature whereas P c (T c ) is the vapor pressure over the condenser sheet. k = h/L vw γ = 0.622 h/C a p is the mass transfer coefficient, where L vw [J kg −1 ] is the specific latent heat of water vaporization, γ is the psychrometric constant, C a is the specific heat capacity of air, and p is the atmospheric air pressure. Here, h = 5.9 + 4.1 u (511 + 294)/(511+T a ) is the heat transfer coefficient, where u and Ta are the prevailing horizontal wind speed and the ambient temperature at 2 m from the ground.  where a is the condenser short-wave albedo (here it is 0.84), S c is the condenser surface area (here it is 1 m 2 ), ε c is the emissivity of the condenser (here it is 0.94), σ is Stephan-Boltzmann constant (5.67 × 10 −8 W m −2 K −4 ), T c [K] is the temperature of the condenser, and R sw and R lw [W m −2 ] are the incoming short-wave radiation (i.e., surface solar radiation downwards) and incoming long-wave radiation (i.e., surface thermal radiation downwards) P cond W Conductive heat exchange between the condenser surface and the ground. For simplicity, we assumed that the condenser is perfectly insulated from the ground; i.e., P cond = 0 P conv W Convective heat exchange P conv = S c (T a − T c ) h where S c is the condenser surface area (here it is 1 m 2 ), T a [K] is the ambient temperature at 2 m from the ground, T c [K] is the temperature of the condenser, and h [W m −2 K −1 ] is the heat transfer coefficient that is estimated based on a semi-empirical equation [37] h = 5.9 + 4.1 WS (511 + 294)/(511 + T a ) and here WS [m s −1 ] is the prevailing horizontal wind speed at 2 m from the ground.
is the specific latent heat of water vaporization and L vi [J kg −1 ] is specific latent heat of water desublimation. Here, dm w /dt is the change rate of water whereas dm i /dt is the change rate of ice In practice, the wettability of the surface affects vapor pressure P c directly above it. In other words, P c is lower over a wet surface; and thus, condensation may take place even if T c > T d (e.g., [15]). According to the model setup, Equation (A3) assumes irreversible condensation; i.e., there is no evaporation or sublimation during daytime even if T c > T a . Furthermore, the model simulation resets the cumulative values for water and ice condensation at noon. and takes the preceding maximum value of m w + m i as the representative daily yield. This way, the model simulation replicates the daily manual dew water collection of the condensed water around sunrise; i.e., after which T c is often above the dew point temperature. available data sources (satellite, radiosondes, aircraft, buoy data, stations, etc.) into a coherent and balanced form of the atmospheric dynamic and thermodynamic state [63]. The ECMWF database is differentiated into 3 main categories: reanalysis, instantaneous forecast, and accumulated forecast. In our case, U10, V10, Ta, and DP were obtained from the reanalysis fields at 00:00, 06:00, 12:00, and 18:00 UTC whereas, at 03:00, 09:00, 15:00, and 21:00 UTC they were instantaneous forecasted fields. Both the forecasted and the reanalysis fields provided 3 hourly time resolution. According to the ECMWF database, the horizontal wind components (U10 and V10) are provided at 10 meters. Therefore, the wind speed at 2 meters was calculated by using a logarithmic wind profile = 2 + 10 + U 10 2 + V 10 2 (A4) where z0 is the surface roughness and U10 and V10 are the horizontal wind speed components at 10 meters. It is important to understand that Equation (A4) is a simple equation that is valid during certain conditions. For instance, in stable conditions (such as at night) Equation (A4) overestimates wind speed at 2 m whereas in unstable conditions Equation (A4) underestimates the wind speed at 2 m. According to the ECMWF database, z0 was obtained as an instanton forecast parameter whereas Rsw and Rlw were accumulated forecasted fields. The mean Rsw and Rlw in a time interval are obtained by taking the difference of the accumulated values between the corresponding time steps divided by the time difference in seconds. The result is a mean value for that time interval 00:00 or 12:00. All input parameters had a horizontal resolution of 0.25 degrees (approximately 30 km).

A.3. Example of a Detailed Model Simulation Output
To show the model output results in detail, we selected Irbid station from 23-28 September 2008 ( Figure A1), which had the maximum dew yield in this period based on the model simulation result. This station is located in the north of Jordan with about 80 km distance from the Mediterranean Sea. Besides the model simulation output, which includes dew yield and temperature of the condenser sheet (Tc). Figure A1 also shows the corresponding model input: long-wave radiation (Rlw), short-wave radiation (Rsw), wind speed (WS), ambient temperature (Ta), and dew point (DP). Here it is clearly shown that condensation occurs when Tc < DP and WS is mostly less than 1 m s −1 . In general, at this station, the dew formation is expected to start in the evening (around 18:00) and Here it is clearly shown that condensation occurs when T c < D P and WS is mostly less than 1 m s −1 .
In general, at this station, the dew formation is expected to start in the evening (around 18:00) and continues to the next morning (until 06:00). The daily collection of dew occurs at noon. During this period, the total cumulative dew yield is~2 mm. This dew yield is equivalent to~2 L collected on a 1 m 2 condenser sheet (i.e., L m −2 ).
During the daytime, the incoming short-wave radiation from the sun as well as the atmospheric long-wave radiation act to increase the temperature of the condenser sheet. In contrast, during dark periods, the outgoing thermal radiation exceeds the atmospheric long-wave radiation. This one is greatly influenced by cloudiness. For example, the thermal emission by clouds, especially low clouds, increases the incoming thermal radiation at the surface.