Harvesting Dew with Radiation Cooled Condensers to Supplement Drinking Water Supply in Semi-arid Coastal Northwest India

This paper describes the development of dew harvesting systems for use in the semi-arid coastal region of northwest India. These systems were developed to ameliorate the drinking water problem in the region, especially for the people living near the coast where groundwater is of poor quality and surface sources are scarce. Although the amount of dew is much less (20-30 mm) than the rains (200-300 mm), it is a more consistent water source. Dew nights number approximately 100 while rainy days number 15-20. Dew occurs over a seven month period (October to April), while rain occurs over a four month period (June September). Although engineered specifically to harvest dew, the designed systems also harvest rain, providing varying amounts of potable water throughout the year. A four-year R&D program led to development of three types of systems condenser-on-roof (CoR), condenser-on-ground (CoG) and Roof-as-Condenser (RaC). The CoR and CoGs systems employ condensers made of plastic film insulated on the underside. CoRs are constructed over the roof of buildings while CoGs are constructed on open ground. The RaCs use the metal roof of buildings as the condenser itself. The CoR and CoGs gave higher output but required higher investment. The RaCs gave lower output but required only a small investment in collection and storage. Examples of working installations are presented. The benefits to the region, the learning accrued, and the partnerships created in the course of work are briefly discussed.


INTRODUCTION
Dew is atmospheric water vapor which condenses on a surface (wind shields, blades of grass) which has been cooled below the dew point temperature of the surrounding air by losing heat to the sky via radiation.Moisture from dew is an important means of survival for plants, arthropods and other organisms in water scarce semi-arid and arid environments 18,24 .However, dew is not considered to be an important source of water for humans because of the small quantities produced and its infrequent occurrence.This is true in general but there are areas where dew occurs frequently and in appreciable quantities which could be of significance to humans.
The potable water shortage in the area is widespread and chronic.More than 150 villages near the coast have no water source and the villagers survive on water hauled daily from long distance by trucks.In such a context dew water could provide a small but critical water source for humans if efficiently harvested.
Monitoring studies in the semi-arid coastal region of northwest India, initiated in 2002, indicated that dew occurred over 103 nights during a season of seven months from October to April 18 (see Figure 1).The study developed a method to collect dew water daily which was Although dew water may have bacteriological contaminants added by air borne dust, the dew water in the study was tested and found to be potable.After filtering (where needed), dew water may be safely pasteurized (Figure 2).Another useful feature of the dew phenomena was that unlike the rains which are concentrated over 15-20 days in a season of four months, and in some years may be negligible, dew occurrence (103 nights) was more uniformly distributed over the season of seven months.Another important note is that the two seasons, rainy and dew, are complementary.

FIGURE 2 DEW WATER FILTERED WITH FINE CLOTH
Condensation occurred frequently on the greenhouse roof because it cooled to the dew point temperature of surrounding air at night by radiating heat to sky.A search of the literature reveals a maximum potential of 0.8 mm/day, or 0.8 litre/m 2 , of dew water can be condensed under radiation cooling 9 .This is based on the available cooling power (25-100 W/m 2) in various regions with respect to the latent heat of condensation (2.26 kJ/g).Peak collection from the greenhouse roof (200 μm of clear polyethylene) on a night was only 0.362 mm or 0.362 liter/m 2 .This indicated that by employing a more efficient radiator material and with proper engineering, a larger amount of dew water could be harvested.The greenhouse roof and the classing were not specifically meant for dew harvesting.
Accordingly, a research and development project was started with the objective of developing efficient and affordable dew harvest devices for the region.The measurements cited above were carried out over months during 2002-03.The project started soon after.It took 3-4 years to gather more measurements, test prototypes, construct pilots and make working devices now available in the region.The project was led by the author from the start, through the various subsequent stages of experimentation, prototyping and field trials.Several research engineers, named in the acknowledgement, worked in the project for various length of time.Senior undergraduate students of engineering did summer internships at Kothara.
The author continues to supervise the work of promotion and the training workmen who install and repair working systems in the area.Work is also ongoing, seeking to improve the design, based on feedback from users and repair teams.Working systems are built on commission for users who may be individuals, businesses and others such as schools.The user-client pays for the installation.Only two small installations were built at small private homes at Kothara, free of cost, for demonstration and awareness raising.Similarly several sets of small condenser units were given free of cost to three high schools to help students learn about the dew phenomena.A team of workmen has been trained to install and later repair the systems.Design of new systems (sizing, configuring) accomplished by the author.This paper is an update containing information on the working installations that have since been added.An overview is provided of the following: (a) a review of the literature (b) a description of the construction of small condenser units to test materials for suitability in order to make large condensers and subsequently the construction of pilot scale units from promising materials, and finally (c) development of working systems for use by the people in the coastal region.

REVIEW OF LITERATURE
Dew has received only sporadic research attention in India with the focus of such research being mainly on assessing its contribution to moisture at the root zone for crops during dry periods. 17; and to the crust of sand dunes in the Thar desert 23 .Efforts to develop devices to condense large amounts of dew water for human use have mostly been reported from Europe.Past efforts, especially those made in the early part of the last century, have been carefully studied by contemporary scientists 13,14 .Results of these studies indicate that past efforts were mostly unsuccessful because the structures erected for the purpose were massive and did not cool down to the dew-point temperature during the night.If there was any dew collected, it was much below expectation.
Peak dew collections in one night from artificial surfaces at various locations are shown in Table 1.Berkowicz et al. 4 measured dew water condensation in Jerusalem using a similar condenser unit as at Kothara.They reported a collection of 33 mm during the 12 month period, over 176 dew events with a maximum in a night of 0.5 mm.Coastal areas of the northwest, India appears to be richer in dew resource than many other locations.Interest in dew condensers has emerged again in the recent years.Current approaches differ fundamentally from the older ones.With a vastly improved understanding of the physics of condensation, efforts have been undertaken to develop light weight, more efficient, condensers made of plastic films 12,15,16,23 .Valuable insights on dew formation, conditions conducive to condensation, materials and substrates that give high yields have emerged from these works.It has been reported that a condensing device made of polyethylene foil pigmented with TiO2 and BaSO 4 (abbreviated PETB here) gave peak recovery of 0.12 liter /m 2 on a night in early tests at Dodoma, Tanzania 16 .As will be seen presently, peak yield from similar foil was much larger in the coastal area of the northwest India.
The attributes of a good condenser and also how to erect it have been studied 14 .These studies indicated that "the 'ideal' condenser should be 'grass-like', i.e. a light sheet thermally isolated from the massive parts and the ground.It is important that the surfaces should be open to let them radiate the energy into space.It means that nothing that can reflect the radiation of the condenser should be placed near its surface and vice versa.The condenser itself should be placed far enough from such surfaces, e.g. the ground, to avoid the 'greenhouse' effect.The material of the sheet should be well wetted by water to reduce the nucleation barrier.Placement of the condenser should be chosen in an open area where the winds are not strong and dew is frequent (i.e.where humidity is high enough).Similar inferences can be drawn by examining equations describing the heat balance on a flat surface out in the open at night, insulated on the underside, with the top facing the sky 19 .For rapid cooling, the surface must have a high emissivity and a low thermal mass (be thin), a view of the sky as fully as possible, and good insulation on the back to prevent heat gain from the ground below.
Radiative cooling, critical to dew deposition, refers to the phenomena of passive cooling under a clear sky.Any surface exposed to a clear sky will emit thermal radiation into the surrounding atmosphere.It will also receive similar radiation emitted by the atmosphere.The surface will steadily cool down at night if the energy emitted is greater than that received and absorbed.If the surface cools to the dew point temperature of the surrounding air, a necessary condition for condensation of moisture from the air is created.The other conditions conducive to condensation are high levels of humidity, a calm environment, and some properties of the surface that facilitate deposition.In a tropical atmosphere, a black surface on the ground could produce up to 71 W/m 2 of useful cooling depending on the difference between the ambient temperature and the temperature of the colder radiating surface.The larger this difference, the smaller the useful cooling power.For a difference of 5 °C, close to 40 W/m 2 cooling power could be expected to be available.
The net outgoing long wave radiation (4.5 to 100 μm) -which is the difference between the outgoing and the incoming -was monitored at the Indian Meteorology Department.It provided a measure of radiation-cooling power available at numerous locations.Data was collected using pyrometers at two time-points -at 05:30 and 20:30 hrs IST.Overall mean net long wave radiation over India lies between 30 to 100 W/m 2 (Mani) 11 .Trivandrum is a coastal location ( 08 0 29 ' N, 76 0 57 ' E, 64 m a.s.l) and is a humid tropical area.There, the cooling power magnitudes are from 38 to 59 w/m 2 .At Jodhpur ( 26 0 18 ' N, 73 0 01 ' E, 224 m a.s.l.), which is a semi-arid region of the northwest, it is above 60 w/m 2 in all months except the rainy season -June to September.On clear nights it is higher, reaching a peak of 72 w/m 2 .Thus there is a fair amount of cooling power available in the area of our interest, the northwest.The cooling power is one of the conditions for dew condensation; availability of highly humid air is another.Both these occur frequently near the coast in northwest region.
Atmospheric characteristics of a region have to be treated as given.It is only the radiator material, the geometry, dimensions, and orientation of the condensers that are under the control of the design engineer.Emission from the sky is low, in the band of 8-13 μm, which is referred to as the 'atmospheric window'.If a radiator is developed which has zero reflectance in this band a colder surface could be achieved. 7That is because in this window the atmosphere is very nearly transparent to thermal radiation.Besides emissivity, some additional properties are also to be considered if the material is to be used to build large practical dew harvest devices.Materials used need to be safe for the collection of potable water; that is, non-toxic.It needs to be durable under long exposure to weather, particularly sun and high temperatures and wind.It should be low cost so as to keep the product water prices competitive.Working systems need to be cost-effective and affordable.It is also desirable that the material be easily available or produced close to the areas of application.

DESCRIPTION OF CONDENSER UNITS
For this project, three readily available materials were initially selected to determine their condensation efficiency.These were (a) PETB film -polyethylene mixed with 5% TiO 2 and 2% BaSO 4 -similar to that used by researchers cited above, (b) galvanized iron (GI) sheet, and (c) an aluminum sheet.A PETB film of the specified composition was obtained from a local plastic processor on order.The other two were purchased from the market.GI sheets are used as roofing on large warehouses and factory sheds.Aluminum sheets too are used externally but less commonly.Both of these are expensive and not expected to be used to build condensers.These were selected only to assess the potential of existing metal roofs for production of water without the use of any other artificial surface.Specifications of the three are given below.
The aim of a year-long empirical investigation was to determine the amount of dew water that would be collected per unit area by each material surface during a season and gain insights into the conditions that are conducive to condensation.A test ground was prepared at the village of Kothara (23° 14 N,68° 45 E, 21 m a.s.l.) with the installation of a multi-channel data logger and sensors to monitor ambient conditions.Small condenser units (1x1 m), similar to those promoted by the International Organization of Dew Utilization (OPUR), were fabricated for the determination of the condensation efficiency of the select materials.Each unit (Figure 3) consisted of three parts -panel (A), frame (C) and collection gutter and accessories (B).The panel was made of two separate sheets laminated together with adhesive.The sheet on top was made of the material to be tested.The underside was insulated with 25 mm thick board of styrene foam.Panels were mounted on the frame at 30° from the horizontal to facilitate draining by gravity and also efficient radiation cooling. 2 Collection accessories (channel and bottle) were also fastened to the frame.The higher end of the panel is 1.5 m from the ground.A total of 12 units were fabricated with four of each of the three materials.Three units (one each of three materials) were installed facing north, three south, three east and three west.All twelve were installed over a test ground at Kothara away from buildings and trees.This ground is well instrumented.Dew water was collected in the plastic bottles and the volumes were noted each morning at 08:00 a.m.The condenser surfaces were not scraped, even though some moisture was easily visible at the time of making the measurement, and which evaporated shortly thereafter.In addition to determining the amount dew water condensed by the three materials, the plan also aimed at determining the effect of orientation on output.surface was measured by a similar RTD element, bonded lengthwise to the surface with thermally conductive adhesive (Product no 34313 Loctite).The element was 3 mm wide and 10 mm long.On occasions, a hand held IR thermometer (from EXTECH Instruments, France) was also used to verify surface temperature.Condenser units were visited at intervals during the night on several occasions to observe the process of condensation.Typically the surfaces would be dry til about 22:00 hours; thereafter tiny dots of moisture would be deposited on it.Gradually these would become more numerous covering the entire surface and also would become larger.Bigger ones would touch the neighbouring droplets and coalesce.Draining of water down the surface could be seen soon after midnight and would continue till the morning hours.On the nights when large amounts of dew condensed, the relative humidity would be upwards of 85% past midnight and the condenser surface would cool below the ambient by 4 o C or more, with calm winds.The dew point temperature shown on the graph was computed using relative humidity values.136

FIGURE 4 COOLING OF CONDENSER SURFACE AT NIGHT
Daily data over a two year period taken from Kothara, 2004 and 2005, was pooled and the means computed.Dew occurred in the area over a span of seven months from -October to April.A small amount of condensation did occur in the last week of September before the onset and the first week of May before the close of the season.It is absent from most of May to most of September, which is the rainy season during which the sky is often covered with clouds to varying degrees and radiation cooling is hampered.There were two peaks -one centered over March -April (summer), the other over October (fall).The quantity of water collected on most (60%) nights varied more or less uniformly between 0.05 and 0.25 mm.The peak condensation collected on a night was 0.55 mm.
Seasonal totals differed to a small extent with the orientation of the measurement units, higher when the units were oriented west and north.For instance, the north and the west oriented units of the PETB condensers collected nearly equal amount -20 mm.The east -oriented unit collected 5% less and the south-oriented unit 15% less.(Table 3) A plausible reason could be that the north and the west-oriented units are not exposed to illumination in the morning as early as others, permitting dew deposition for a little longer.Also for the same reason, east and south faced units may be losing more by evaporation.A more significant difference, however, was apparent when examining the material of construction.Column (6) of Table 3 shows the average collection from all four units of the three materials.The highest collection was in the PETB units (19.4 mm) followed by GI (15.6 mm) and aluminum (9 mm).The number of nights the dew collected in the PETB and the GI units were equal, 114-115.On aluminum units condensation occurred on significantly fewer (85) nights.

Pilot Scale condenser Units
Aluminum was dropped at this stage due to poor performance results.The PETB and corrugated GI sheet were used to build pilot units.The PETB pilot was built over the ground.It was made of two 3x3 m modules placed side by side oriented east providing a total condensing area of 18 m 2 ( Figure 5).The condensing surface with the PETB film on top was insulated on the underside in the same manner as the smaller ones described above.The surface was built over a sand bed inclined 30 o from the horizontal, similar to the smaller units mounted on frames.The GI pilot also was of 18 m 2, 22 but mounted on a gable frame, one half oriented west and the other east (Figure 6).No insulation was installed underneath the GI sheets in order to mimic the existing buildings which normally are un-insulated.The pilot was 4 m from the ground at the ridge as the GI sheds usually are.The GI sheet was newly purchased and was 1.5 mm thick with a shiny surface and IR emissivity measured as 0.23.Both halves were pitched at 30 o .There were separate gutters and collection gear for each side.Dew water collection was measured daily over an entire season.The PETB pilot unit yielded 15 mm over the season (Table 4), approximately 20% lower than that of the smaller PETB unit oriented east (19.7 mm -Table 3).The yield from GI pilots was 11 mm from west and 9.98 mm from east.These values were also lower than the corresponding small units (13.7 west, 11.5 east).In the case of the GI pilot, the main reason for the difference would appear to be lack of insulation.However the output is significant enough to encourage owners of the large GI roofed properties to use the roof to harvest dew water.Also they could be advised to improve output by installing a layer of insulation under the roof.At conclusion of the pilot tests it was decided that when new systems are built the PETB film would be employed.

EXAMPLES OF WORKING INSTALLATIONS OF DEW HARVEST SYSTEMS
For marketing purposes, models of working systems were conceptualized.Simultaneously a campaign was carried out in the area to inform people living in the coastal villages about the dew harvest technology and to remove misgivings some held about dew water.Some residents of these villages held the view that consuming dew water is harmful for humans as well as for cattle.Open-house get-togethers were organized at the campus of the Development and Outreach Station, Kothara (the base of our research projects).The gettogethers were open to all, but formal invites were also posted to farmers, school teachers, government officers dealing with water problems and non-government-organizations operating in the area.Presentations were made to explain the working of dew harvest systems.Dew water quality, chemical and microbial test reports from accredited laboratory were explained.The amount of water that can be harvested in a season, as well as the costs of installations were discussed.Although normal water was served to visitors, dew water was also on offer and project staff made it a point to use it in these events.These events were held twice a year.Gradually the apprehensions dissolved and enquiries began to be received for installations.The installations have three components (a) the condenser surface (b) the water collection and conveyance accessories, and (c) storage.For new installations, two types were conceptualized -one with the condenser surface built over roof of buildings, termed condenser on roof (CoR), and the other built over open ground or condenser on ground (CoG).Both types would have a PETB condenser.At the same time, owners of large GI roof buildings were to be advised to use the roof itself as a condenser termed roof as condenser (RaC).

Roof-as-Condenser type system at Suthari
The first commission to install a RaC type system was received from a temple complex at Suthari, a village 2 km from the coast of the Arabian Sea.Local ground water was not potable.Surface sources dried out soon after October, the last month of rainy season.The community lives on water hauled daily on tanker-trucks from a source several kilometers away.The complex had several GI roofed warehouses to store hay and offered one of the 343 m 2 areas for dew harvesting (Figure 7).The roof was gabled with one half facing west and the other east, both pitched 15 o .Being the first installation of this type it was instrumented.A temperature sensor was installed on the upper side of the roof duly shielded and bonded to surface by thermally conductive adhesive.A five channel data logger was installed to monitor roof temperature, ambient temperature, wind speed and direction and relative humidity.Table 5 and Figure 8 show the cooling of the roof on a night.At 1600 hrs in the afternoon the roof was hotter than the ambient air by nearly  6).This could be due to lack of insulation and due to the pitch not being steep enough 20 .Although the yield was low, it was obtained at only a small investment in collection gear ($250).Water from new installations was tested in accredited laboratory at Ahmedabad at least two to tree times in the initial months after start of regular use.Findings of one of the tests are shown in Table 8.Water is generally potable.In one of the subsequent tests water showed some bacteriological contamination; that is, E Coli which may have come from air borne dust from the surrounding fields.Human excreta and cattle dropping are found in the fields.More over there was a large cattle barn right next to the warehouse.Water was rendered safe by boiling it.The initial investment was $250.The fittings were made of UV stabilized PVC and were expected to last over ten years.No maintenance was required, barring occasional cleaning of dust from the roofing sheet.Taking the total seasonal dew yield to be 1500 L, and a discount rate of 10 %, calculations showed that the investment will be recovered in ten years if water is valued at 4 cents/L.Processed bottled water is available in the market for 20 cents/L.This is of course not generally used by people.Bulk water in 20 L returnable carboys is available for 3 cents/L.
Considering the dew yield alone, the system is competitive with the bulk water rates.In addition to the dew water, the 343 m 2 will harvest rain water.Taking 100 mm as the high probability rainfall in the season, a total of 34,000 L of water would be collected.Dew harvest surfaces are very efficient means of rain harvest making the systems economically more attractive.
This RaC was visited by a large number of people, being part of a temple complex.It provided an opportunity for visitors to taste dew water which further helped in removing their apprehensions.

Condenser-on-Roof type system at Sayara
This CoR was commissioned by a school in Sayara, a village 15 km from the coast of Arabian Sea and not far from Suthari just described.Indeed the school teachers had visited there and sampled dew water over several visits.The system was installed on roofs of its three buildings. 21All roofs were pitched 15°, one half facing north, and other south, and were made of reinforced concrete with total area of 360 m 2 (Figure 9).Styrene foam boards of 1x1 m size were bonded to roof surface with hot bitumen.PETB film (0.2 mm thick) was fixed over the surface anchored firmly by 15 mm wide UV stabilized plastic ribbon.The ribbonends were screwed into the walls.Dew water flowed to the gutters and into the collection tank just below ground level.Total yield of dew water during 2005-06 was 3622 liters or 10 mm (Table 7).Findings from one of the tests of dew water are shown in Table 7.Note that the yield from PETB pilot unit was 15 mm (Table 4).But again the roof pitch at Sayara was less steep than the pilot.The condenser on the roof top made the buildings cooler inside by reducing the heat gain from the roof.
The cost of installation was approximately $1,000.The plastic film was replaced every three years ($240); and other fixtures are expected to last over ten tears.No maintenance is required barring occasional cleaning of dust from the condensation surface.Taking the total seasonal dew yield to be 3600 L, and a discount rate of 10%, calculations showed that the investment will be recovered in ten years if water is valued at 6 cents/L.The expected yield of rain water from 360 m 2 of surface area was 36,000 L. The school has a 20,000 L underground covered storage for the purpose.

Condenser-on-Ground type systems at Panandhro and Satapar
A CoG type of system at Panandhro was commissioned by the Gujarat Mineral Development Corporation, a large mining company (Figure 10).Before building, dew was measured for a season using small condensers described earlier.A total of 16 mm was collected and there was only a small variation with orientation.The site was judged suitable for dew harvest.Also the land is of low fertility.Accordingly it was decided to build a CoG type system 5 .The condenser field consists of an array of ten ridge-and-trough modules built next to each other.Ridges, each 35 m long, were first formed by bucket excavators over gently sloping ground.Ridges are oriented along the slope of the ground.Each ridge is trapezoidal (top 50 cm, base 200 cm, two sides sloping 30° from horizontal, height 100 cm).The ridge is lined with PETB

LEARNING
When in 2002 the possibility of harvesting dew water for human use was visualized, there was little reliable information available on the phenomena in this region.The scientists (meteorologists, engineers) held the view that it was too little to bother.People of villages near the coast (where it frequently occurred) held the belief that consuming dew water was injurious to health.There was little prospect of getting funding support for research.To move forward it was decided to make actual measurements (amount and frequency) of dewfall by collecting dew condensation from greenhouse roof, get the water tested in accredited laboratories and publish the findings.An experimental greenhouse already existed at the Development and Outreach Station (DOS) Kothara; so no extra funds were needed to get measurements going.Publication of results was helpful in improving the attitudes of engineers and scientists.It helped in getting a small grant to make more systematic

FIGURE 1
FIGURE 1 SEMI-ARID COASTAL REGION OF NORTHWEST INDIA (Source: MapPoint)

FIGURE 3 CONDENSERS
FIGURE 3 CONDENSERS TEST GROUND AT KOTHARAAmbient conditions at the site (air temperature, relative humidity, wind speed) were recorded continuously by a data logger (Weather Technologies India, Pune).(Table2, Figure4) It had an 8 channel LCD display, real time clock calendar, serial output port (RS 232C) for connection to a computer and was provided with software for data retrieval from memory module.The temperature sensors were standard Platinum RTD elements (PT 1000), model DWT 8102 mounted inside a weather shield coated with weather proof reflective white paint.The sensor resolution was 0.1 °C, with an accuracy of ± 0.2 °C, within the temperature range of -40 °C to + 60 °C.The relative humidity sensor was solid state capacitive (DTH 8103) type with an accuracy of ± 3 % of full scale reading (resolution ± 0.1 %) and a range of 0-99%.Wind speed was measured by a three cup anemometer installed 6 m above ground.The anemometer assembly was mounted on a friction free shaft coupled to an 18 slot chopper.It had a starting threshold of 0.3 m/s, a range of 0-65 m/s, and an accuracy of better than ± 0.5 m/s.These were also provided by the same company.The temperature of the condenser FIGURE 5 PILOT PETB CONDENSER AT KOTHARA FIGURE 8 COOLING OF GI ROOF -RaC SUTHARI

TABLE 2
The dew yield on the night of 17 th and 18 th was 290 ml on PETB unit

TABLE 3
11 o C. It began to cool rapidly after the sun set and became cooler than the ambient by 1900 hrs.It continued to cool reaching 4 o C below the ambient by 2200 hrs and remained cooler than the air till the morning.Relative humidity was 90 % and above after midnight, with winds calm.It is in these conditions that condensation occurred.total of 30 liters was collected by the next morning.In the year 2005, this RaC received 1497 liters of dew over 96 nights, in 2006, 1581 liters over 85 nights.The mean seasonal water yield was 4.5 mm (Table A

TABLE 7 DEW
WATER COLLECTION -CoR SAYARA

Table 8
Indian Standards (ISO 10500 1993) for Drinking Water and Quality of Dew Water at Three Installations Findings of one of the water test are shown in Table8.The school normally treats the stored water with chlorine tablets.The total cost of the installation was $2500.All CoGs and CoRs also harvest rain water in rainy season.