Movatterモバイル変換

CHAPTER 7 - ACTIVE PROTECTIONMETHODS

Active protection methods include activities that are doneduring a frost night to mitigate the effects of subzero temperatures. Thesemethods include:

1 Heaters
2 Wind machines
3 Helicopters
4Sprinklers
5 Surface irrigation
6 Foam insulation
7 Foggers
8Combinations of active methods

The cost of each method varies depending on local availabilityand prices. For example, a range of costs for commonly used systems is given inTable 7.1. However, the benefits sometimes depend on multiple uses of the system(e.g. sprinklers can also be used for irrigation). The costs and benefits ofselecting a particular system are discussed in Volume II, Chapter 2 on the"Economic evaluation of protection methods." The theory of operation, propermanagement and the advantages and disadvantages of each of the active protectionmethods are discussed in this chapter.

TABLE 7. 1
The required number of protection devices perhectare and a range of estimated costs in US dollars per hectare for the year2000 for installation and operation in deciduous orchards and vineyards inWashington State (USA) (R.G. Evans, pers. comm.)

Heaters

One method to replace the losses of energy from a crop, in afrost situation, is to compensate with the massive use of fuel (solid, liquid orgas) burnt in heaters of various types. Depending on orientation of the heatersrelative to the plants, part of the radiation is directly intercepted by plantparts, which raises the plant temperature. In addition, air that is heated bythe fire is transported by free and, if the wind is blowing or wind machines areused in combination, forced convection to the plants and air within and abovethe canopy. Weather conditions that favour efficiency of this method are calmconditions with little or no wind and the presence of a stronginversion.

Heaters have been used to protect crops from freezing for atleast 2000 years and the effects and methodology are well known. Generally, theheaters fall into two categories. There are heaters that raise the temperatureof metal objects (e.g. stack heaters) and there are those that operate as openfires. Protecting with heaters is technically dependable and growers preferredheaters until pollution problems and high costs of fuel relative to the cropvalue made the method too expensive for many crops. Now heaters are mainly usedto supplement other methods during extreme frost events and for high-valuecrops. In this section, the following topics are discussed:

• Theory ofoperation

• Smoke effects

• Heater requirements

• Heater placement andmanagement

• Liquid-fuel heaters

• Propane-fuel and naturalgas-fuel heaters

• Solid-fuel heaters

• Mobile heaters

Theory of operation

Natural energy losses from a crop are bigger than the gainsduring a frost night and this causes the temperature to drop. Energy is mainlylost to net radiation and the losses are partially replaced by sensible and soilheat fluxes towards the surface (Figure 7.1). If condensation (i.e. dew orfrost) occurs, then released latent heat can also replace some of the energyloss. Heaters provide supplemental energy (Q) to help replace the net loss(Figure 7.1). If sufficient heat is added to the crop volume so that all of thelosses are replaced, the temperature will not fall. However, there isinefficiency in the operation of heaters and, under some conditions, it becomescost prohibitive to introduce sufficient energy to make up for the systeminefficiency. Proper design and management can improve the efficiency to thelevel where the crop is protected under most radiation frost conditions.However, when there is little or no inversion and the wind is blowing, theheaters may not provide adequate protection.

Heaters provide frost protection by direct radiation to theplants around them and by causing convective mixing of air within the inversionlayer (Figure 7.2).

FIGURE 7.1
An orchard in an imaginary box, where theenergy fluxes represented are net radiation (R_n), vertical andhorizontal sensible heat flux (H),conductive heat flux from theground (G), latent heat (LE) and energy added by heating(Q)

FIGURE 7.2
Hot air rises and cools until about the sameas the ambient temperature, then it spreads out and cools until it becomesdenser and descends; this creates a circulation pattern

Most of the energy from heaters is released as hot gases andheated air that mainly warms the ambient air by convection. Radiant energy fromthe heaters travels directly to nearby plants that are in direct line-of-sightof the heaters. However, depending on the crop canopy density and structure,only a small percentage of the radiant energy from stack heaters isintercepted.

The energy requirement to prevent damage during a radiationfrost is roughly equal to the net radiation loss (e.g. between -90 Wm^-2 and -50 W m^-2), minus downward sensible heat flux andupward soil heat flux. Both the sensible and soil heat flux densities varydepending on local conditions, but it is likely that 20 to 40 W m^-2are contributed by each source. Therefore, the energy requirement to preventfrost damage is in the range 20 to 40 W m^-2. Heater energy output istypically in the range 140 to 280 W m^-2; depending on the fuel,burning rate and number of heaters. Therefore, much of the energy output fromheaters is lost and does not contribute to warming the air or plants and theefficiency, which is defined as the energy requirement divided by the energyoutput, tends to be low. However, proper management can increase efficiency ofthe energy supplied by heaters.

The air temperature leaving a stack heater is between 635°C and 1000 °C, so the less dense heated air will rise rapidly afterleaving a heater. As the heated air rises, because of entrainment with coldersurrounding air, expansion of the heated air parcels and radiation, it coolsrapidly until it reaches the height where the ambient air has about the sametemperature. Then the air spreads out, mixing with other air aloft. Eventually,the mixed air will cool, becomes denser and descend, which creates a circulationpattern within the inversion layer (Figure 7.2). If the inversion is weak or ifthe fires are too big and hot, the heated air rises too high and a circulationpattern within the inversion is not produced. Modern heaters have more controlover the temperature of emitted gases to reduce buoyancy losses and improveefficiency. The most efficient systems have little flame above the stack and nosmoke. Operating the heaters at too high a temperature will also reduce thelifetime of the heaters.

When there is a strong inversion (i.e. a low ceiling), theheated air rises to a lower height and the volume influenced by the heaters issmaller. Because the heated volume is smaller, heaters are more effective atraising the air temperature under strong inversions. Heater operation is lessefficient at increasing air temperature in weak inversion (i.e. high ceiling)conditions because they have a bigger volume to heat. Under weak inversionconditions, using a fuel with a higher fraction of energy output to radiationthan to heating the air will improve protection. This fraction is commonlyimproved by having more and smaller heaters, with exhaust funnels that retainheat. Also, when fires are too big or hot, the warmed air can break through thetop of the inversion, there is less circulation in the inversion layer and theheaters are less efficient at warming the air (Figure 7.3).

Because heaters warm the air, the air inside a protected cropis generally rising and cold air outside is being drawn in from the edges toreplace the lifted air. Consequently, more frost damage occurs and hence moreheaters are needed on the borders. Kepner (1951) reported on the importance ofinversion strength and placing more heaters on borders. He studied a 6.0 hacitrus orchard that was warmed with 112 chimney heaters burning 2.8 litreh^-1 with an average consumption of 315 l ha^-1h^-1. The unprotected minimum air temperature was 1.7 °C, but theresults are similar to what one expects on a radiation frost night. The orchardwas square and the easterly wind varied from 0.7 m s^-1 to 0.9 ms^-1(2.5 km h^-1 to 3.2 km h^-1). Figure 7.4shows how the temperature varied in a transect across the centre of the orchard.The wind direction was from the left. The upper graph (A) shows the effects ofheater operation on temperature during two nights with differing inversionstrength. The lower graph (B) shows the benefits from using twice the number ofheaters on the upwind border.

FIGURE 7.3
Diagram of a frost night temperature profileand the influence of heater output on heat distribution and loss from anorchard

FIGURE 7.4
Temperature effects of heater operation (A)under differing inversion conditions and (B) with different concentrations ofheaters on the upwind border (Kepner, 1951)

The 6.0 ha citrus orchard had trees that were about 4.6 mtall and 4.6 m diameter planted on either 6.7 × 6.7 m or 6.1 × 7.3 mcentres. Heaters were placed in the tree rows with one heater per two treeswithin the orchard and one heater per tree on the upwind border when theconcentration was increased. The orchard was warmed with about 112 chimneyheaters burning 2.8 l h^-1with an average consumption of 315 lha^-1 h^-l. The wind direction was from theleft.

In Figure 7.4, the increase in temperature was highest midwayacross the orchard and the benefits from heating were less near the upwind anddownwind borders. On the night with 4.2 °C of inversion strength, theincrease in temperature on the upwind edge was about 40 percent of the increasein the middle of the orchard (Figure 7.4.A). The increase in temperature on thedownwind edge was about 60 percent of the increase in the middle of the orchard.On the night with 7.8 °C of inversion strength, the temperature atmid-orchard was about 1.0 °C warmer than on the night with 4.2 °Cinversion strength (Figure 7.4.A). The wind speed was slightly higher during thenight with 7.8 °C inversion strength, so the difference most probablyresulted from more efficient use of the convective heat within the strongerinversion layer.

In Figure 7.4.B, the temperature was increased by nearly 1°C on the upwind edge when there was one heater per tree rather than oneheater per two trees along the upwind border. There was less benefit fromadditional heaters on the downwind border, but, because the wind direction mightchange, it is wise to place additional heaters on all borders. Edge effects areimportant and well known by growers. In fact, growers will at times extinguishsome fires when heaters are lit in neighbouring orchards.

Smoke effects

Today, it is well known that the protection from heaters comesfrom the heat released by the fires and not from smoke production (Collomb,1966). Smoke does cover the sky and reduces visibility, but it has negligibleeffect on the apparent sky temperature. The dimension of the average smokeparticle is less than 1.0 mm diameter (Mee andBartholic, 1979), which reduces radiation in the visible range (0.4-0.7 mm) but has little effect on transmission of long-waveradiation. Therefore, upward long-wave radiation from the surface mainly passesthrough the smoke without being absorbed. Consequently, smoke has little effecton upward or downward long-wave radiation at night and hence has little benefitfor frost protection. Because smoke offers little or no benefit and it pollutesthe air, it is better to minimize smoke production and maximize thermalefficiency of the combustion. Smoke at sunrise blocks solar radiation and delaysheating of the crop, which can lead to higher fuel consumption and possibly moredamage. There are reports that gradual thawing of frozen citrus reduces damage(Bagdonas, Georg and Gerber, 1978), but there are other reports that indicatethere is no evidence for this belief (Burkeet al.,1977). If true, thensmoke might be beneficial, but modern pollution laws make the use of smokeillegal in most locations. Where orchards are small and close to roads, heatersmoke has been known to cause automobile accidents, as in northern Italy, whichled to serious legal and insurance problems. Consequently, smoke generation isnot recommended for frost protection.

Heater requirements

Liquid-fuel heaters typically provide about 38 MJ of energyper litre of fuel and the output energy requirement varies between 140 and 280 Wm^-2 (5.0 and 10 GJ ha^-1h^-1) depending on thefrost night conditions (Blanc et al., 1963). Dividing the energy requirement inJ ha^-1h^-1 by the energy output J l^-1, the fuelrequirement varies between 133 and 265 litre ha^-1h^-1. Thenumber of burners needed depends on the desired level of protection and theburning rate of the heaters. If each heater consumes about 1.0 litreh^-1, then dividing the fuel requirement by the consumption rate givesa range between 133 and 265 liquid-fuel heaters per hectare(H_H).For more efficient protection, it is best to keepthe fuel consumption per heater low and use more heaters.

The energy output for commonly used liquid and solid fuels isprovided in Table 7.2. Note that the energy output is in MJ l^-1 forliquid-fuel, MJ per cubic metre for gas and MJ per kilogram for solid fuels. Ifthe fuel consumption rate (F_C)and energy requirement(E_R),including additional energy required forinefficiency, are known, then the number of heaters per hectare can bedetermined. Use Equation 7.1 to determine the number of liquid-fuel heaters perhectare from the energy requirement (E_R)in Wm^-2, the fuel energy output (E_O)in MJl^-1 and the fuel consumption rate (F_C)inlitre h^-1 per heater:

The 3.6 × 10⁷ coefficient convertsE_Rin W m^-2 to J h^-1 ha^-1.For solid fuels, use Equation 7.1 to determine the number of heaters per hectare(H_H)from the energy requirement(E_R)in W m^-2, the fuel energy output(E_O)in MJ kg^-1 and the fuel consumptionrate (F_C)in kg h^-1 per heater. An Excelapplication program "HeatReq.xls" for calculating both liquid-fuel andsolid-fuel heater requirements is included on the computer applicationdisk.

Heater placement andmanagement

Heater distribution should be relatively uniform with moreheaters in the borders, especially upwind and in low spots (Figure 7.5). If thecrop is located on a slope, then more heaters should be placed on the upslopeedge where cold air is draining into the crop. Under freezing conditions, whenthe wind speed exceeds 2.2 m s^-1 (7.9 km h^-1),considerable heat loss occurs due to horizontal advection and higherconcentrations of heaters are needed on the upwind border. Low spots, which arecolder, should also have higher concentrations of heaters. Heaters on theborders should be lit first and then light more heaters as the need increases(e.g. if the wind speed increases or the temperature drops). Heaters areexpensive to operate, so they are commonly used in combination with windmachines or as border heat in combination with sprinklers.

TABLE 7.2
Energy output for a variety of commonfuels

NOTE: (1) Output depends on wood type, watercontent of fuel, and size and number of fires. Energy outputs are expressed inMJ per litre, MJ m³ or MJ kg^-1 for liquid, gas and solidfuels.

FIGURE 7.5
Sample arrangement of heaters (small dots inthe figure), with higher concentrations along the upwind edge and in low spots(after Ballard and Proebsting, 1978).

Liquid-fuel heaters

Liquid-fuel heaters were developed for frost protection duringthe early 1900s. Use of the method decreased as oil prices and concerns aboutair pollution increased. Although not widely used, the use of liquid-fuelheaters for frost protection is still a viable method in cases where laws do notprohibit it and the cost of fuel is not too high. Liquid-fuel heaters requireconsiderable labour for placement, fuelling and cleaning, in addition to thecapital costs for the heaters and the fuel. Typically, there are about 75 to 100oil stack heaters or 150 to 175 propane-fuel heaters per hectare, and a welldesigned and operated heater system will produce about 1.23 MW ha^-1(i.e. 123 W m^-2) of power. The approximate consumption rate is 2.8litre h^-1 per heater for oil- and kerosene-fuelled heaters and about1 m³ h^-1 for propane-fuel heaters. More than half of theenergy output from the heaters is lost as radiation to the sky and convectiveheat losses on a typical radiation frost night, so the heater output is highrelative to the heat gained by the crop. Note that these recommendations are forprotection of large deciduous orchards that are surrounded by other orchardsthat are being protected. Isolated smaller orchards may require moreheaters.

When lighting heaters, every second or third heater in a rowshould be lit first. Then go back and light the remaining heaters. This helps toreduce convective losses of heat through the top of the inversion layer.Oil-fuel heaters should be cleaned after every 20 to 30 hours of operation, andthe heaters should be closed to prevent entry of rainwater that could causeleakage of oil onto the ground. The stack can be blown off or the fireextinguished if too much steam is produced. Remove oil from the heaters at theend of the season. Free-flame-type heaters will accumulate carbon and lower thefuel efficiency level. Catalytic sprays can be used to reduce carbonaccumulation. They should be refilled before they run out of fuel and cleanedwith a stick or simply hit to free the soot accumulation that reducesefficiency.

Various types of fuels burners are available for frostprotection. A list of fuels and heaters approved for use in Florida (USA) aregiven in Tables 7.3 and 7.4. Because they can be improvised with cans of paint,oil, etc., the free flame type (i.e. without a chimney) is cheaper and easier totransport and fill. They are smaller, so the density of heaters can be greater,giving better mixing and less chance for the chimney effect. This sometimesresults in improved protection. However, they are less fuel-efficient becausethere is more volatilization and they pollute more. In some locations, they arenot approved for use in frost protection.

TABLE 7.3
Liquid-fuel and gas fuels approved by theFlorida Department of Environmental Protection for frostprotection

TABLE 7.4
Heaters approved by the Florida Department ofEnvironment for frost protection

Air pollution regulations are often quite stringent and localregulations should be reviewed before purchasing or using heaters. Most regionalauthorities have similar regulations on burning fuels for frost protection. Inaddition, some authorities have other requirements for use of heaters. Forexample, the Florida State Environmental Agency requires that, when usingheaters for frost protection, air temperature be measured using a standardlouvered weather shelter or fruit frost shelter (Figures 6.1 and 6.2). All localregulations should be investigated before using heaters.

An equal mixture of fuel and gasoline [petrol] is used tolight heaters. Buckets or tanks towed by a tractor, which allow filling of twolines of burners simultaneously, are used to refill the heaters after a frost.When direct heating is used, to minimize fuel consumption the protection isstarted just before reaching critical damage temperatures. The temperatureshould be measured in a Stevenson screen, fruit-frost shelter or Gill shieldthat shields the thermometers from the clear sky.

Propane-fuel and natural gas-fuelheaters

Labour requirements to refill liquid-fuel heaters are high, sosome growers moved from using individual heaters to centralized distributionsystems. The systems use tubing to transport the fuel to the heaters. The fuelcan be natural gas, liquid propane or fuel oil. In more elaborate systems,ignition, the combustion rate and closure are automated in addition to fueldistribution. The capital cost to install centralized systems is high, but theoperational costs are low. Propane-fuel heaters require less cleaning and theburning rates are easier to control than oil-fuel heaters. Because the burningrate is less, more heaters are needed (e.g. about 130-150 per hectare), but theprotection is better. Under severe conditions, the propane supply tank cansometimes freeze up, so a vaporizer should be installed to prevent the gas linefrom freezing.

Solid-fuel heaters

Solid fuels were used as a method for frost protection beforeliquid or gas fuels. As liquid fuels dropped in price, there was a switch fromsolid to liquid fuels, especially in North America. When it was discovered thatthe ratio of radiation to total energy released was about 40 percent for burningsolid fuels (e.g. wood, coal and coke) in comparison with 25 percent for burningliquid fuels (Kepner, 1951), there was a revival in the use of solid fuels.Having a higher ratio of radiation to total energy release is important asconditions become windier (e.g. during an advection frost). The maindisadvantage of solid fuels is that the energy release diminishes as the fuel isused up, and energy release thus becomes limiting when needed most (Hensz,1969a; Martsolf, 1979b). Another drawback is that solid fuels are difficult tolight, so they must be started early. They are also difficult to extinguish, sofuel is often wasted if started when unnecessary.

A variety of solid fuels are used for frost protection (e.g.wood, coke, old rubber tyres, paraffin candles and coal). Some oil companiesmarket products consisting of petroleum wax - a refinement by-product - and cokethat appear in various forms, including candles and bricks.

When compared with the liquid-fuel burners, solid fuels oftenshow better results. For example, using two oil wax candles under eachgrapefruit tree in an orchard gave an average increase of 1.7 °C to thefruit. Energy efficiency (i.e. the fruit temperature inside and outside of thefoliage) from using the wax candles was more than double that of liquid-fuelburners (Miller, Turrell and Austin, 1966). An increase of 2.2 °C at 1.1 mheight was observed from using 375 bricks of petroleum wax and coke per hectare(Parsons, Schultz and Lider, 1967). Conventional liquid-fuel burners requiretwice the energy output to gain the same effect on air temperature in thecanopy. For example, petroleum wax heaters used only 60 percent of the energynormally needed to get the same protection (Schultz, Lider and Parsons, 1968).Modification of temperature within the inversion layer was more concentratednear the ground - where the crop is - when burning petroleum wax and coke brickscompared with feedback chimney burners (Gerber, 1969). Thus, to improveefficiency it is clearly better to have many small fires than a few bigfires.

Mobile heaters

A mobile heater is commercially available as a method forfrost protection; however, scientific evaluations of the machine have not yetbeen published. The mobile heater uses four 45-kg propane tanks to supply thefuel for the heater, which mounts on the back of a tractor (Figure 7.6). Theheater uses a centrifugal fan to blow the heated air horizontally andperpendicular to the tractor direction as it moves up and down the rows. Afterstarting the heater, the fuel supply is adjusted to give a temperature ofapproximately 100 °C where the air vents from the machine. When operated,the airflow extends to 50 to 75 m either side of the machine. The tractor isdriven up and down rows far enough apart so that the area of influence overlaps.The manufacturer recommend that the tractor make one complete cycle through thecrop about every 10 minutes, a period allowing coverage of about 5-7ha.

In some unpublished experiments, the mobile heater showedlittle effect on the minimum temperatures recorded within protected orchards.Since the energy output from the machine is much less than energy losses from acrop during a radiation frost night, this was not unexpected. However, wheneverthe machine passed by a point within the crop, there was a short-lived increasein the temperature recorded with exposed thermocouples. It is possible thatthese short-lived temperature increases have a positive effect and preventfreezing of the plant tissue; however, more research is needed to verify if thisis true.

FIGURE 7.6
A mobile heater for frost protection mountedon the back of a tractor

Photo: R.L. Snyder

Recently, some researchers have suggested that the mobileheater might be beneficial because it dries the plant surfaces. Since watertypically freezes on the outside of plant tissue and then propagates inside thetissue to cause freezing in intercellular spaces, there may be some validity tothis theory. However, more research is clearly needed to validate effectivenessof the machine.

Wind machines

Conventional windmachines

Wind machines (or fans) that blow air almost horizontally wereintroduced as a method for frost protection in California during the 1920s.However, they were not widely accepted until the 1940s and 1950s. Now they arecommonly used in many parts of the world. Wind machines are used on a widevariety of crops including grapevines, deciduous trees and citrus. Californiacitrus orchards are nearly all protected by wind machines.

Wind machines generally consist of a steel tower with a largerotating fan near the top. There is usually a two- or four-blade fan with adiameter typically varying from 3 to 6 m. The typical height for fans is about10-11 m above ground level. However, lower heights are used for lower canopies.To our knowledge, the fan height is set to avoid hitting the trees and there isno aerodynamic reason for the height selection. The most effective wind machineshave propeller speeds of about 590 to 600 rpm. Fans rotate around the tower withone revolution every four to five minutes. Most wind machine fans blow at aslight downward angle (e.g. about 7 °) in the tower direction, whichimproves their effectiveness. When the fan operates, it draws air from aloft andblows at a slightly downward angle towards the tower and the ground. Power tooperate the fan usually comes from an engine mounted at the base of the tower;however, some of the older machines have engines that rotate with the fan at thetop of the tower. Matching the rotation of fans around their towers so that allfans are blowing in the same direction is believed to improve mixingeffectiveness.

Before investing in wind machines, be sure to investigate thelocal climate and local expenses. For example, if there is little or noinversion, then wind machines are not recommended. In California, wind machinesare widely used in citrus orchards, which are mainly protected during Decemberthrough January, but not in deciduous orchards, because inversions tend to bestrong during winter months when citrus need protection, but not in the CentralValley in the spring when deciduous trees need protection. There are reportsthat wind machines work better after deciduous trees leaf out in the spring.Consequently, fans are not often used in almond orchards that commonly needprotection before leaf-out. Conducting a temperature survey to measuretemperature inversions during the frost protection period before purchasing windmachines is advisable. If there is little or no inversion, then select adifferent protection method. Locate machines in places where the wind drift isenhanced by the fans. In some instances, installing machines where they can pushcold air out of low spots can be beneficial.

Wind machines generally have lower labour requirements andoperational costs than other methods. This is especially true for electric windmachines. However, when electric wind machines are installed, the grower isrequired to pay the power company "standby" charges, which cover the cost ofline installation and maintenance. The standby charges are paid whether the windmachines are used or not. In fact, because of increases in the cost of power,electric wind machines have become marginally cost-effective for citrusprotection in some regions of California (Venner and Blank, 1995). Internalcombustion wind machines are more cost-effective because they do not have thestandby charges. However, they require more labour. The capital cost to installwind machines is similar to sprinkler systems, but operational costs arehigher.

Generally, except for noise, wind machines are environmentalfriendly. Wind machine noise is a big problem for growers with crops near citiesand towns. This should also be considered when selecting a frost protectionmethod.

Theory of operation

Wind machines provide protection by increasing the downwardsensible heat flux density and by breaking up microscale boundary layers overthe plant surfaces. Fans do not produce heat, but redistribute sensible heatthat is already present in the air. The fans mix warm air aloft with colder airnear the surface (Figure 7.7). They also help by removing the coldest air closeto the leaves and replace it with slightly warmer ambient air. The amount ofprotection afforded depends mainly on the unprotected inversion strength. Theinversion strength is calculated as the difference between the 10 m and 1.5 mtemperatures in an unprotected orchard. Within the region influenced by a windmachine, the average air temperature measured at 1.5 m increases by about 1/3 ofthe inversion strength. Near the wind machine tower, the protection achieved isbetter (Figure 7.8). The actual benefit depends on inversion characteristics,which cannot be generalized. However, stronger inversions clearly give betterprotection.

FIGURE 7.7
A schematic diagram that shows the effect ofwind machines on the temperature profile during a radiation frost

FIGURE 7.8
Traces of 10 m and 2 m temperature trendswithout (w/o) a fan and 2 m temperature trends measured near wind machinesstarted at 0145 and 0315 h

Temperature measurements were collected 30 maway from the wind machine

FIGURE 7.9
Temperature field response to wind machine on26 March 2000 in northern Portugal

(A) Temperature profiles (30 m from windmachine) before and after wind machine; and (B) 1.5 m temperature responsepattern produced by wind machine after 2 complete rotations around the tower(after Ribeiro et al., 2002).

Generally, a 75-kW wind machine is necessary for each 4 to 5ha (i.e. a radius of about 120 m to 125 m). If one wind machine is used, about18.8 kW of engine shaft power per hectare is typically needed. About 15 kW ofengine shaft power is suggested per machine per hectare when several machinesare used. Protection decreases with distance from the tower, so some overlap ofprotection areas will enhance protection. Usually, the protection area is anoval rather than a circle shape because of wind drift. For example, Figure 7.9Bshows the 1.5 m height temperature response pattern to wind machine operation inan apple orchard (Ribeiroet al.,2002).

Starting and stopping

Wind machines are typically started when the air temperaturereaches about 0 °C. Under stable inversion conditions, air tends tostratify near the ground and mixing is believed to become less. However, trialsin California (USA) and Portugal have shown that starting fans after inversionshave formed has little influence on fan effectiveness. In less than half-an-hourafter starting, the 2 m temperature typically rises, sometimes approaching the10 m temperature of an unprotected orchard (Figure 7.8). However, because thetemperature of a radiating surface during a frost night is usually lower thanthe air temperature, it is wise to have the wind machines operating when the airtemperature reaches the critical damage temperature (T_c). Ifthe fruit is wet during the day or evening of an expected frost night, the windmachines (and heaters) should be started earlier to attempt to dry the fruitbefore ice can form on the fruit. Wind machines are not recommended when thewind is blowing at more than about 2.5 m s^-1 (8 km h^-1) orwhen there is supercooled fog. When the wind is more than 2.5 m s^-1,it is unlikely that an inversion is present and it is possible that the fanblades could experience wind damage. A simple method to estimate the wind speedis to hang plastic ribbon from a tree branch or street sign. Police crime scenetape is an example of the type of plastic ribbon that could be used. Such ribboncan be purchased from farm or engineering supply stores. If the wind is greaterthan 2.5 m s^-1, the bottom of the hanging ribbon should be blowingback and forth to about 30 cm from horizontal.

In a supercooled fog, water droplets can freeze on the fan andsevere wind machine damage can occur if the ice cases one blade to break off butnot the other.

Vertical flow windmachines

The use of vertical flow fans to pull down the warm air alofthas been investigated; however, these fans generally work poorly becausemechanical turbulence mixing with the trees reduces the area affected by theventilation. Also, the high wind speed near the base of the tower can damagehorticultural and ornamental crops. Wind machines that blow vertically upwardsare commercially available and there has been some testing of the machines. Theidea is that the fan will pull in cold dense air near the ground and blow itupwards where it can mix with warmer air aloft. In theory, cold air is removednear the surface and the warmer air aloft drops downward hence lowering theinversion. Limited testing has shown that this method has a temporary positiveeffect on temperatures near the fan; however, the extent of influence andduration of the effect is still unknown.

To our knowledge, the method has only been used in smallvalleys where cold air ejected upwards is likely to fall back towards thesurface. In a location where prevailing winds aloft might move the airhorizontally away from the crop, more protection could result. However, there isno known research evidence at this time.

Helicopters

Helicopters move warm air from aloft in an inversion to thesurface. If there is little or no inversion, helicopters are ineffective. Due tothe large standby and operational costs, the use of helicopters for frostprotection is limited to high value crops or emergencies (e.g. when the normalmethod breaks down).

Authors differ on their estimates of the protection area forhelicopters. The area covered by a single helicopter depends on the helicoptersize and weight and on the weather conditions. Estimated coverage area variesbetween 22 and 44 hectares (Evans, 2000; Powell and Himelrick, 2000). Passes areneeded every 30 to 60 minutes, with more passes under severe conditions. Waitingtoo long between passes allows the plants to supercool and the agitation from apassing helicopter can cause heterogeneous nucleation and lead to severedamage.

Temperature sensors are often mounted on the outside of thehelicopter and the pilots fly at a height where they observe the highesttemperature reading. The optimal height is commonly between 20 and 30 m. Commonflight speeds are 25 to 40 km h^-1 (Powell and Himelrick, 2000) or 8to 16 km h^-1 (Evans, 2000). Higher velocities have not improvedprotection. Temperature increases between 3.0 °C and 4.5 °C are commonfor a hovering helicopter. Pilots often load helicopter spray tanks with waterto increase the weight and increase thrust. Under severe frosts with a highinversion, one helicopter can fly above another to enhance the downward heattransfer.

Thermostat controlled lights at the top of the canopy are usedto help pilots see where passes are needed. As the helicopter passes over thecrop, the temperature rises and the lights go out. Cooling to the thermostattemperature causes the lights come on. This helps the pilot to find and fly overcold spots. Alternatively, a ground crew should monitor temperature in the cropand communicate with the pilot where flights are needed.

On the sides of hills, heat transfer propagates down-slopeafter reaching the surface. Therefore, flying over the upslope side of a cropusually provides more protection because the effects are felt downwind as well.Flights are stopped when the air temperature upwind from the crop has risenabove the critical damage temperature.

Sprinklers

Using sprinklers for frost protection has the advantage overother methods that water application is generally less expensive. The energyconsumption is considerably less than that used in frost protection with heaters(Gerber and Martsolf, 1979) and, therefore, operational costs are low comparedto heaters and even wind machines. Labour is mainly needed to ensure that thesystem does not stop and the heads do not ice up during the night. In additionto frost protection, one can use the sprinklers for irrigation, enhancing fruitcolour by over-plant evaporative cooling, reducing sun injury by over-plantirrigation, delaying bloom prior to bud break, fertilizer application and acombination these applications. Also, the method is relatively non-polluting.The main disadvantage of using sprinklers is the high installation cost andlarge amounts of water that are needed. In many instances, a lack of wateravailability limits the use of sprinklers. In other cases, excessive use canlead to soil waterlogging, which could cause root problems as well as inhibitcultivation and other management activities. Nutrient leaching (mainly nitrogen)is a problem where sprinkler use is frequent. In some instances, excessive useof sprinklers can affect bacterial activity in the soil and it can delaymaturation of fruit or nuts (Blancet al.,1963). In this section on useof sprinklers, the following topics are discussed.

1 Basic concepts
2 Over-plantsprinklers

Conventional rotatingsprinklers
Starting and stopping
Application raterequirements
Variable-rate sprinklers
Low-volume targetedsprinklers

3 Sprinkling over coverings
4 Under-plantsprinklers

Conventional rotatingsprinklers
Microsprinklers
Low-volume (trickle-drip) irrigation
Heatedwater

Basic concepts

Like air, water has sensible heat that we measure with athermometer and the water temperature increases or decreases depending onchanges in the sensible heat content. When water temperature drops, it happensbecause (1) sensible heat in the water is transferred to its surroundings, (2)water vaporizes, which consumes sensible heat to break the hydrogen bondsbetween water molecules, or (3) there is net radiation loss. As water dropletsfly from a sprinkler head to the plant and soil surfaces, some sensible heat islost to radiation, some will transfer from the warmer water to the cooler airand some will be lost to latent heat as water evaporates from the droplets. Theamount of evaporation is difficult to estimate because it depends on the watertemperature and quality, droplet size and path length and weatherconditions.

Understanding changes in sensible heat content of water andconversions between sensible and latent heat are crucial to understand frostprotection with sprinklers. Water temperature is a measure of the sensible heatcontent of the water and heat released to the air and plants as the watertemperature falls provides some of the protection. From wells, water commonlyhas a temperature near the mean annual air temperature at the location. For thewater temperature to fall from 20 °C to 0 °C, each kg (or litre) mustlose 83.7 kJ of sensible heat. This heat can be lost by radiation; transferredto sensible heat in the air, plants or ground; or it can contribute toevaporation. When 1.0 kg of water freezes at 0 °C, the phase changeconverts 334.5 kJ of latent heat to sensible heat. The total amount of energyreleased in cooling water from 20 °C and freezing it is 418.3 kJk^-1. If the initial water temperature were 30 °C rather than 20°C, then cooling to 0 °C would provide an additional 41.9 kJkg^-1 for a total of 460.1 kJ kg^-1. However, cooling 1.1 kgof 20 °C water and freezing it provides 460.9 kJ k^-1, soapplying 10 percent more water provides the same energy as heating the water by10 °C.

The cooling and freezing of water replaces energy lost duringa radiation frost. However, evaporation from the surface removes sensible heatand causes the air temperature to drop. Although evaporation rates are low, theenergy losses can be high. For a phase change from liquid to water vapour (i.e.evaporation), the loss is 2501 kJ kg^-1 at a temperature of 0 °C.For a phase change from ice to water vapour (i.e. sublimation), the loss is2825.5 kJ kg^-1 at 0 °C. Therefore, the energy released bycooling 1.0 kg of 20 °C water to 0 °C and freezing it is 418.3 kJkg^-1, and the amount of water cooled and frozen must be more than 6.0times the amount evaporated (or 6.8 times the amount sublimated) just to breakeven. Additional energy from the cooling and freezing process is needed tocompensate for the net energy losses that would occur withoutprotection.

When water droplets strike a flower, bud or small fruit, thewater will freeze and release latent heat, which temporarily raises the planttemperature. However, energy is lost as latent heat when water vaporizes fromthe ice-coated plant tissue. This, in conjunction with radiation losses, causesthe temperature to drop until the sprinklers rotate and hit the plant withanother pulse of water. The secret to protection with conventional over-plantsprinklers is to re-apply water frequently at a sufficient application rate toprevent the plant tissue temperature from falling too low between pulses ofwater. For non-rotating, low-volume, over-plant, targeted sprinklers, the ideais to continuously apply water at a lower application rate, but targeted to asmaller surface area.

For conventional under-plant sprinklers, the idea is applywater at a frequency and application rate that maintains the ground surfacetemperature near 0 °C. This increases long-wave radiation and sensible heattransfer to the plants relative to an unprotected crop. For under-plantmicrosprinklers, which apply less water than conventional sprinklers, the goalis to keep only the ground under the plants near 0 °C, to concentrate andenhance radiation and sensible heat transfer upwards into the plants.

Over-plant sprinklers

Over-plant sprinkler irrigation is used to protect low-growingcrops and some deciduous fruit trees, but not for crops with weak scaffoldbranches (e.g. almond trees), where excessive weight of ice on plants could snapbranches. It is rarely used on subtropical trees (e.g. citrus) except for younglemons, which are more flexible. Even during advection frosts, over-plantsprinkling provides excellent frost protection down to near -7 °C if theapplication rates are sufficient and the application is uniform. Under windyconditions or when the air temperature falls so low that the application rate isinadequate to supply more heat than is lost to evaporation, the method can causemore damage than would be experienced by an unprotected crop. Drawbacks to thismethod are that severe damage can occur if the sprinkler system fails, themethod has large water requirements, ice loading can cause damage, and rootdisease can be a problem in poorly drained soils.

Application rate requirements for over-plant sprinklers differfor conventional rotating, variable rate and low-volume targeted sprinklers. Inaddition, the precipitation rate depends on (1) wind speed, (2) unprotectedminimum temperature, (3) the surface area of the crop to be covered and (4)distribution uniformity of the sprinkler system. As long as there is aliquid-ice mixture on the plants, with water dripping off the icicles, thecoated plant parts will maintain their temperature at about 0 °C. However,if an inadequate precipitation rate is used or if the rotation interval of thesprinklers is too long, all of the water can freeze and the temperature of thecoated plants will again start to fall.

Conventional rotatingsprinklers

Conventional over-plant sprinkler systems use standard impactsprinklers to completely wet the plants and soil of a crop. Larger plants havemore surface area, so a higher application rate is needed for tall than forshort plants. For over-plant sprinklers to be effective, the plant parts must becoated with water and re-wetted every 30 to 60 seconds. Longer rotationintervals require higher application rates.

Sprinkler distribution uniformity is important to avoidinadequate coverage, which might result in damage. A sprinkler system evaluation(i.e. a catch-can test) should be performed prior to frost season to be surethat the application uniformity is good. Information on how to perform asprinkler can test is usually discussed in most textbooks on irrigationmanagement, and guidelines are often available from local extension advisors. Ifcold air is known to drift in from a specific direction, increasing sprinklerdensity on the upwind edge of the crop, or even in an open field upwind from thecrop, can improve protection. Do not include the higher density area in thesystem evaluation area.

Any over-plant irrigation system that delivers an appropriateapplication rate can be used for frost protection, but systems specificallydesigned for frost protection are best (Rogers and Modlibowska, 1961; Raposo,1979). The system needs to be in place during the entire frost season. Once inplace and operating during a frost night, a system cannot be moved. Generally,the distribution uniformity is improved by using an equilateral triangle ratherthan rectangular head spacing. Systems designed for irrigation rather than frostprotection can be used providing uniformity is good and the precipitation rateis adequate. In most cases, the sprinkler heads should be mounted at 0.3 m orhigher above the top of the plant canopy to prevent the plants blocking thespray. For frost protection, specially designed springs, which are protected byan enclosure to prevent icing of the heads, are typically used. Clean filtersare needed to be sure that the system operates properly, especially when riveror lagoon water is used.

Portable hand-move sprinkler systems with the heads risingjust above the canopy top can be used for low growing crops like strawberries.For deciduous trees and vines, use permanent sprinkler systems with eithergalvanized or polyvinyl chloride (PVC) pipe risers that place the heads justabove the canopy top. Wooden posts can support the risers. Typical sprinklerhead pressures are 380 to 420 kPa with less than 10 percent variation.

Starting and stopping

Starting and stopping sprinklers for frost protection dependson the temperature and humidity in the orchard. When a sprinkler system is firststarted, the air temperature will drop; however, the air temperature will notdrop below the temperature of the water droplets and it will normally rise againonce water begins to freeze and release latent heat.

The effect of over-plant rotating sprinkler application isillustrated in Figure 7.10, which shows the response of leaf-edge temperature towetting by sprinklers every 120, 60 or 30 seconds (based on Wheaton and Kidder,1964). Between wettings, evaporation (or sublimation) occurs and the phasechange from liquid or ice to water vapour converts sensible to latent heat. Theremoval of sensible heat causes temperature of wet plant tissue to fall. Becausethe plant tissue is wet, the temperature will fall to as low as the wet-bulbtemperature. If the dew-point temperature (humidity) is low, then the wet-bulbtemperature can be considerably lower than air temperature, so the temperatureof wet plant tissue can fall well below air temperature and cause damage ifinsufficient water is applied. Also, if the rotation rate is too slow or if thesprinklers are stopped too early, temperatures can drop below the criticaldamage temperature and cause damage.

In older literature on the use of over-plant sprinklers, itwas common to warn against a sharp drop in temperature when sprinklers arestarted during low-dew-point conditions. Under windy low-dew-point temperatureconditions when the air temperature is relatively high (e.g. 10 to 15 °C),evaporation from the droplets causes the water droplet temperature and hence theair temperature to drop rapidly when the sprinklers are started. However, thewater droplet temperature commonly falls to near 0 °C as they pass from thesprinkler heads to the plant surfaces. Consequently, there is no reason why theair temperature would drop below 0 °C when the sprinklers are started. Asshown in Figure 7.10, the plant surface temperatures will drop below 0 °Cas water sublimates from the plant surfaces. But the temperature rises againwhen hit with a new droplet of liquid water.

FIGURE 7.10
Leaf-edge temperature changes when wetted bya sprinkler system applying water at 2.8 mm h^-1 with rotation ratesof 120, 60 and 30 s, air temperature near 0 °C, a wet-bulb temperature near-2 °C and a wind speed near 5.5 km h^-1(after Wheaton andKidder, 1964)

Because the critical damage (T_c)temperatures are somewhat questionable and because they are based ontemperatures of dry rather than wet plants, it may be wise to start thesprinklers when the wet-bulb temperature is a bit higher thanT_c.Starting sprinklers when the wet-bulb temperature reaches 0 °C is lessrisky and it may be prudent if there is no water shortage and waterlogging andice loading are not a problem.

Even if the sun is shining on the plants and the airtemperature is above 0 °C, sprinklers should not be turned off unless thewet-bulb temperature upwind from the crop is aboveT_c.If soilwaterlogging or shortage are not problems, permitting the wet-bulb temperatureto exceed 0 °C before turning off the sprinklers adds an extra measure ofsafety.

The wet-bulb temperature can be measured directly with apsychrometer or it can be derived from the dew-point and the air temperature.For direct measurements, the cotton wick on the wet-bulb thermometer is wettedwith distilled or de-ionized water and it is ventilated until the temperature ofthe wet-bulb thermometer stabilizes. Ventilation is accomplished by swinging asling psychrometer or by blowing air with an electric fan using an aspiratedpsychrometer (Figure 3.9). If the recorded temperature is below 0 °C, thewater on the wick might be frozen. Then the observed temperature is called the"frost-bulb" rather than wet-bulb temperature. However, there is littledifference between the frost-bulb and wet-bulb temperature in the rangeimportant for frost protection.

Rather than using a sling or aspirated psychrometer with aventilated wet-bulb thermometer, a simple thermometer with a wetted cotton wickand no ventilation can be employed to approximate the wet-bulb temperature.However, if possible, it is better to ventilate the cotton wick with a fan. Onecan use for the wick a cotton shoestring that fits snugly on the thermometerbulb.

If you decide to not measure the wet-bulb upwind from thecrop, an alternative is to use the dew-point temperature from a weather serviceor from a measurement. Dew-point sensors are expensive, but a simple method isto use a shiny can, water, salt and ice (Figure 7.11). First pour some saltedwater into the shiny can. Then start adding ice cubes to the can while stirringthe mixture with the thermometer. Watch the outside of the can to see when dewcondenses (or ice deposits) on the surface. If there is no condensation ordeposition, add more ice and salt to further lower the temperature. When you seeice deposit, immediately read the thermometer temperature. The thermometerrecording is the "ice point" temperature, which is a bit higher than, but closeto, the dew-point temperature. Shining a flashlight (torch) onto the can surfacewill help you to see the ice form and to read the thermometer. This method isless accurate than using a dew-point hygrometer, but it is often sufficientlyaccurate for determining start and stop temperature for sprinklers.

FIGURE 7.11
A simple method to estimate the dew-pointtemperature

Slowly add ice cubes to the water in a shiny can to lowerthe can temperature. Stir the water with a thermometer while adding the icecubes to ensure the same can and water temperature.

When condensation occurs on the outside of the can, notethe dew point temperature.

In most of the literature on using sprinklers for frostprotection, the start and stop air temperatures are determined relative to thedew-point and wet-bulb temperatures. In reality, they should be based on the icepoint and frost-bulb temperatures since ice covered plants are more common thanwater covered plants at subzero temperatures. However, a table of airtemperatures corresponding to subzero dew-point and wet-bulb temperatures isnearly identical to a table of air temperatures corresponding to the ice pointand frost-bulb temperatures. Therefore, only the dew-point and wet-bulbtemperatures are used in Table 7.5, to avoid confusion with commonpractice.

To use Table 7.5, locate the wet-bulb(T_w)temperature in the top row that is greater than orequal to the critical damage (T_c) temperature for the crop.Then locate the dew-point (T_d)temperature in theleft-hand column and find the air temperature in the table that corresponds.Make sure that the sprinklers are operating before the air temperature measuredupwind from the crop falls to the selected air temperature. The values in Table7.5 are for sea level, but they are reasonably accurate up to about 500 melevation. For more accuracy at higher elevations, the SST.xls applicationprogram, which is included with this book, does these calculations and it can beused to determine exact starting and stopping temperatures for any inputelevation.

TABLE 7.5
Minimum starting and stopping air temperatures(°C) for frost protection with sprinklers as a function of wet-bulb anddew-point temperature (°C) at mean sea level

NOTE: Select a wet-bulb temperature that is abovethe critical damage temperature for your crop and locate the appropriate column.Then choose the row with the correct dew-point temperature and read thecorresponding air temperature from the table to turn your sprinklers on or off.This table is for mean sea level, which should be reasonably accurate up toabout 500 m elevation.

TABLE 7.6
Dew-point temperature (°C) correspondingto air temperature and relative humidity*

NOTE: Select a relative humidity in the leftcolumn and an air temperature from the top row. Then find the correspondingdew-point temperature in the table.

When using a frost alarm,set the alarm about 1 °Chigher than the starting air temperature identified in Table 7.5 to ensuresufficient time to start the sprinklers. If the sprinkler starting is automatedwith a thermostat, the starting temperature should be set 1 °C or 2 °Chigher than the starting air temperature from Table 7.5, depending on thermostataccuracy.

If only the relative humidity and air temperature areavailable, then use Table 7.6 to estimate the dew-point temperature. Then usethe dew-point temperature and the selected wet-bulb temperature corresponding tothe critical damage temperature to decide the air temperature to start and stopsprinklers.

For those who prefer to use equations to estimate the startand stop air temperatures, the vapour pressure (e_din kPa) atthe dew-point temperature (T_din °C) is estimated fromthe wet bulb temperature (T_win °C) as:

where the saturation vapour pressure at the wet-bulbtemperature is:

and the barometric pressure (P_b)as afunction of elevation (E_L)in metres is:

Therefore, the corresponding air temperature(T_a)can be calculated as:

where the saturation vapour pressure at the dew-pointtemperature is:

Application rate requirements

Application rate requirements for over-plant sprinkling withconventional sprinklers depend on the rotation rate, wind speed and unprotectedminimum temperature. When the wind speed is higher, there is more evaporation,higher sensible heat losses from the plant surfaces and more water must befrozen to compensate for these losses. When the unprotected minimum temperatureis lower, then more energy from the freezing process is needed to make up forthe sensible heat deficit. Sprinkler rotation rates are important because thetemperature of wet plant parts rises when the water freezes, but it falls aswater vaporizes and radiative losses continue between pulses of water strikingthe plants (Figure 7.10).

Frequent wetting of the crop is needed to reduce the timeinterval when the plant temperature falls below 0 °C (Figure 7.10).Generally, the rotation rate should not be longer than 60 seconds and 30 secondsis better. For example, the widely used sprinkler application raterecommendations for grapevines for wind speeds of 0.0 to 0.5 m s^-1(Table 7.7)and wind speeds of 0.9 to 1.4 m s^-1 (Table 7.8)depend on the sprinkler rotation rate and minimum temperature as well as thewind speed. Gerber and Martsolf (1979) presented a theoretical model foroverhead sprinkler application rate for protection of a 20 mm diameter treeleaf.

Using that model a simple empirical equation giving nearly thesame sprinkler application rate (R_A)is givenby:

whereu(m s^-1)is the wind speed andT_l (°C) is the temperature of a dry unprotectedleaf.

Using the approach outlined by Campbell and Norman (1998), thedifference between air and leaf temperature of a 0.02 m diameter leaf on atypical frost night, with high stomatal resistance, can be estimatedas:

for 0.1 £u£5 m s^-1.Combining the twoequations, a simple equation for the sprinkler application rate in terms of windspeed (u) in m s^-1 and air temperature (T_a) in°C is given by:

For practical purposes, the wind speed entered into equation7.9 should fall between 0.5 and 5 m s^-1.An additionalapplication amount should be added to the result from Equation 7.10 to ensuregood wetting of the leaves. The additional amount varies from approximately 0 mmh^-1 for sprinkler systems with uniform coverage over a sparse cropcanopy to 2.0 mm h^-1 for canopies with dense foliage and/or with lessuniform sprinkler coverage.

The application rates generated by Equation 7.9, with thecorrections to ensure adequate wetting, fall in the range of application ratesrecommended for tall crops in Figure 7.12. The values from Tables 7.7 and 7.8for tall grapevines also fall within the range of application rates in Figure7.12. Application rates are less for short crops because there is less surfacearea to cover, there tends to be less evaporation and it is easier to obtainuniform wetting of the vegetation when it is shorter (Figure 7.13). The rates inFigure 7.13 are typical for strawberries and slightly higher rates are appliedto potatoes and tomatoes. Other intermediate sized crops have rates betweenthose shown in Figures 7.12 and 7.13.

TABLE 7.7
Application rates for overhead sprinklersprotection of grapevines based on minimum temperature and rotation rate for windspeeds of 0.0 to 0.5 m s^-1 (after Schultz and Lider,1968)

TABLE 7.8
Application rates for overhead sprinklersprotection of grapevines based on minimum temperature and rotation rate for windspeeds of 0.9 to 1.4 m s^-1 (after Schultz and Lider,1968)

FIGURE 7.12
Tall Crops

Over-plant conventional sprinkler application raterequirements for frost protection of tall crops with head rotation rates of 30 s(horizontal hatching) and 60 s (vertical hatching). Wind speed ranges from 0.0 ms^-1 at the bottom to 2.5 m s^-1 at the top.

FIGURE 7.13
Short Crops

Over-plant conventional sprinkler application raterequirements for frost protection of short crops with head rotation rates of 30s (horizontal hatching) and 60 s (vertical hatching). Wind speed ranges from 0.0m s^-1 at the bottom to 2.5 m s^-1 at the top.

The effectiveness of sprinklers depends mainly on theevaporation rate, which is strongly influenced by wind speed. However, theminimum temperature is an indication of a deficit of sensible heat in the air,so a higher application rate is also needed if the minimum temperature is low.If there is a clear liquid-ice mixture coating the plants and water is drippingoff the ice, then the application rate is sufficient to prevent damage. If allof the water freezes and it has a milky white appearance like rime ice, then theapplication rate is too low for the weather conditions. If the application rateis insufficient to adequately cover all of the foliage, then damage can occur onplant parts that are not adequately wetted. For example, trees could sufferlittle damage on upper branches while damage occurs on lower branches where thebuds, blossoms, fruit or nuts were not adequately wetted. As conditions worsen,more damage will occur. Under windy, high evaporation conditions, inadequateapplication rates can cause more damage than if the sprinklers are notused.

Variable rate sprinklers

For most growers, the selection of a sprinkler precipitationrate is made once and it cannot be easily changed after the system is installed.Most systems are designed to apply the amount needed for the worst conditions inthe region. This leads to over-application on nights when conditions are lesssevere. To overcome this problem some growers design systems with changeableriser heads to permit higher or lower application rates. In addition, variablerate sprinklers, which are switched off and on, have been studied extensively(Gerber and Martsolf, 1979; Proebsting, 1975; Hamer, 1980) as a method to reduceapplication rates. For example, by using an automated variable rate sprinklersystem, Hamer (1980) obtained efficient protection during a frost night usingonly half of the normal amount of water. Water was applied whenever thetemperature of an electronic sensor placed in the orchard to mimic a plant budfell to -1°C. However, he noted that, due to non-uniform application,placement of the temperature sensor was critical. Also, at the end of longperiods of frost protection, ice accumulation slowed the temperature response ofthe sensor and led to excessive water applications. Kalmaet al.(1992)noted that, rather than measuring the temperature of an ice-coated sensor, NZAEI(1987) pulsed sprinklers for one minute on and one minute off whenever theminimum temperature measured by an exposed sensor at 1.0 m in an unprotectedarea was above -2.0 °C and operated the sprinklers continuously for lowertemperatures. This resulted in 18 percent water savings during one season. Amodel to predict the application requirements for variable rate (i.e. pulsed)sprinkler systems is reported in Kalmaet al.(1992).

A recent paper by Kocet al.(2000) reported that up to75 percent water savings was achieved by cycling water off and on with solenoidsduring over-tree sprinkling for frost protection of an apple orchard.Environmental parameters and bud temperatures were used to model the on and offperiods.

Low-volume (targeted)sprinklers

Use of one over-plant micro-sprinkler per tree was reported toprovide good protection with less water use in the southeastern USA (Powell andHimelrick, 2000). However, they noted that the installation costs are high andthe method has not been widely accepted by growers. Evans (2000) reported thatone over-plant microsprinkler per tree will reduce the application raterequirement from between 3.8 and 4.6 mm h^-1 for conventionalsprinklers to between 2.8 and 3.1 mm h^-1for the surface area coveredby trees. However, under windy conditions, rates as high as 5.6 mmh^-1 might be needed to protect orchards.

Jorgensenet al.(1996) investigated the use oftargeted over-plant microsprinklers for frost protection of grape vineyards.They evaluated a pulsing action that produces large diameter droplet sizes whilemaintaining lower application rates, compared with those found in conventionalmicrosprinkler designs. The microsprinkler applies a band of water approximately0.6 m wide targeted over the cordon of the vine. Microsprinklers were installedin every vine row about half a metre above the cordon on every second trellisstake, with approximately 3.6 m between heads. The targeted system was comparedto a conventional impact sprinkler system with 15.6 m by 12.8 m spacing using2.78 mm diameter nozzles. The targeted system had an 80 percent water savings;however, there were no severe frost events during the two-yearexperiment.

Targeted sprinklers were used to protect grapevines at ahigher elevation site (820 m) in northern California (USA) and the results werepromising. In that location, there was a shortage of water, which forced thegrower to look for an alternative to conventional over-plant sprinklers. Thelow-volume system applied approximately 140 litre min^-1ha^-1 as compared with the grower's conventional system application of515 to 560 litre min^-1 ha^-1. In the first year of thetrial, the lowest temperature observed was -3.9 °C, but no difference incrop loads or pruning weights between the low-volume and the conventionalprotection blocks were observed. In the second year, air temperatures fell aslow as -5.8 °C on one night, which was low enough for some of the impactsprinkler heads to freeze up and stop turning. Although there was considerableice loading, the grower observed that the frost damage losses were similar inboth the conventional and low-volume sprinkler blocks. The low-volume sprinklersystem was designed to spray water directly onto the vine rows and little wasapplied to the ground between rows. The grower pointed out that it was importantto orient the non-rotating sprinkler heads to obtain a uniform coverage of thevine rows. It was also important to start the sprinklers when the wet-bulbtemperature was above 0 °C and not to stop until the wet-bulb temperaturerose above 0 °C again.

Sprinkling overcoverings

Sprinkling over covered crops in greenhouses and framesprovides considerable protection. Like over-plant sprinkling, continuousapplication of sufficient water to plant covers keeps the covers at near 0°C. The thin layer of water intercepts upward terrestrial radiation andradiates downward at a temperature near 0 °C, which is considerably higherthan the apparent sky temperature. As a result the net radiation on the plantcanopy is considerably higher than a canopy exposed to the clear sky. Hogg(1964), in a two-year trial, reported average protection of 2.4 °C usingsprinkling irrigation over a Dutch frame (i.e. with a glass cover). Duringcolder nights, the protection was closer to 4.5 °C. However, theprecipitation rate of 7.3 mm h^-1 was high. The use of sprinklers ongreenhouses with 0.2 mm thick plastic covers maintained temperatures inside upto 7.1 °C higher than the subzero temperatures that were registered outside(Pergola, Ranieri and Grassotti, 1983). Relative to an identical plasticgreenhouse that was heated to the same temperature difference, the sprinklerssaved up to 80 percent in energy costs. The sprinklers operated intermittentlyand the average precipitation rates on the colder nights were near 10 mmh^-1, which is a high rate. However, more research is needed todetermine if lower precipitation rates are possible and to study the effects ofwater quality on the plastic. Because there is less surface area to cover, theprecipitation rate should be similar or possibly less than that used over tallcrop canopies. However, this needs further study. The use of sprinklers overplastic greenhouses has also been used in southern Portugal with positiveresults (Abreu, 1985).

Under-plant sprinklers

Under-tree sprinklers are commonly used for frost protectionof deciduous tree crops in regions where the minimum temperatures are not toolow and only a few degrees of protection are needed. In addition to the lowoperational cost, one can also use the system for irrigation, with fewer diseaseproblems and lower cost, so it has several advantages relative to over-plantsprinklers. Also, limb breakage due to ice loading and sprinkler system failureare not a problem with under-plant sprinkler systems. Lower application rates(2.0 to 3.0 mm h^-1) are needed for under-plant sprinkler systems. Theprotection afforded depends on severity of the frost night and the applicationrate. For example, Anconelliet al.(2002) found little benefitdifference between application rates and sprinkler head types for minimumtemperatures above -3 °C. Below -3 °C, an outflow of 65 litreh^-1 per tree gave better performance than 45 litreh^-1.

When under-plant sprinklers are used, the main goal is tomaintain the wetted surface temperature at near 0 °C. Protection derivespartly from increased radiation from the liquid-ice covered surface, which iswarmer than an unprotected surface. In an unprotected orchard, the airtemperature is generally coldest (i.e. often well below 0 °C) near thesurface and increases with height. Because sprinkler operation increases thesurface temperature to near 0 °C, the air near the surface is also warmerthan in an unprotected crop. The warmed air near the surface creates atmosphericinstability near the ground and causes upward sensible heat flux to warm the airand plants. In addition, the water vapour content of air in the orchard isincreased by the sprinkler operation and condensation or deposition of ice onthe cold plant surfaces will release some latent heat and provideprotection.

The effectiveness of the sprinklers again depends on theevaporation rate, which increases with wind speed. The best way to test yoursystem is to operate the sprinklers during various freezing conditions duringdormancy to identify conditions when all of the water freezes. If the soil iscovered with a liquid-ice mixture and the surface temperature is at 0 °C,the application rate is adequate. If all the water freezes and the surfacetemperature falls below 0 °C, then the application rate is too low forthose conditions. Care should be taken to avoid wetting the lower branches ofthe trees.

Conventional rotatingsprinklers

Perry (1994) suggested that temperature rises of between 0.5°C and 1.7 °C up to a height of about 3.6 m are expected during atypical radiation frost when using rotating under-plant sprinklers. Evans (2000)indicates that temperature increases up to about 1.7 °C are possible at 2.0m height in an orchard protected with cold water. Connell and Snyder (1988)reported an increase of about 2 °C at 2.0 m height in an almond orchardprotected with a gear-driven rotating sprinkler head system rather than impactsprinklers. Water temperature from the sprinkler heads was about 20 °C andthe application rate was 2.0 mm h^-1. Typical under-plant sprinklersystems use 2.0 to 2.4 mm diameter, low-trajectory sprinkler heads with 276 to345 kPa of pressure and application rates between 2.0 and 3.0 mmh^-1.

Once started, the sprinklers should be operated continuouslywithout sequencing. If water supply is limited, irrigate the areas most prone tofrost or areas upwind from unprotected orchards. It is better to concentratewater on areas needing more protection than to apply too little water over alarger area. Good application uniformity improves protection. Hand-movesprinkler systems should not be stopped and moved during a frost night. However,under mild frost conditions (T_n > -2.0 °C), thesprinkler lines can be placed in every second row rather than every row to covera larger area. For moderate to severe frosts, closer spacing of the sprinklerlines may be necessary.

Several researchers (Perry, 1994) have recommended that havinga cover crop is beneficial for protection when under-tree sprinklers are usedfor frost protection. This recommendation is based partially on the idea thatthe presence of a cover crop provides more surface area for water to freeze uponand hence more heat will be released (Perry, 1994; Evans, 2000) and partly onthe idea that the height of the liquid ice mixture and hence the height wherethe surface temperature is maintained at 0 °C is elevated closer to thetree flowers, buds, or fruits that are being protected (Rossiet al.,2002). The difficulty in having a cover crop is that, although there mightbe additional protection if and when the system is used, it is also more likelythat active protection will be needed if a cover crop is present. Where waterand energy resources are limited and frosts are infrequent, it might be wiser toremove the cover crop and reduce the need for active protection. In climateswhere frosts are common and there are adequate resources to operate theunder-plant sprinklers, then maintaining a cover crop may improve protection.However, energy and water usage will increase.

Microsprinklers

In recent years, under-plant microsprinklers have becomeincreasingly popular with growers for irrigation and interest in their use forfrost protection has followed. Rieger, Davies and Jackson (1986) reported on theuse of microsprinklers with 38, 57 and 87 litre h^-1 per treeapplication rates and two spray patterns (90° and 360°) for frostprotection of 2-year old citrus tree trunks that were also wrapped with foilbacked fibreglass insulation. The trees were spaced at 4.6 × 6.2 m, so theequivalent application rates were 218, 328 and 500 litre min^-1ha^-1 or 1.3, 2.0 and 3.0 mm h^-1. On a night when thetemperature fell to -12 °C, the trunks of trees in the irrigated treatmentswere 1.0 to 5.0 °C higher than non-irrigated control temperatures. Thetemperature difference between the 57 and 87 litre h^-1 applicationrates was insignificant, but the trunk temperatures were somewhat higher thanfor the 38 litre h^-1 application rate. However, even the trunktemperatures for the 38 litre h^-1 treatment fell only to -2.5 °Cwhen the air temperature fell to -12 °C, so clearly the combination ofmicrosprinklers with trunk wraps was beneficial. The authors also reported thata 90° spray pattern gave better protection than the 360° pattern.There was no measurable difference between air temperatures or humidity in theirrigated and non-irrigated treatments, but the upward long-wave radiation washigher in the irrigated plots.

More protection is afforded by covering a larger area withwater; however, there is additional benefit coming from water placed under theplants where radiation and convection are more beneficial than water placedbetween crop rows. However, if you spread the same amount of water over a largerarea, the ice is likely to cool more than if the water is concentrated into asmaller area. Again, the best practice is to supply sufficient water to cover aslarge of an area as possible and be sure that there is a liquid ice mixture overthe surface under the worst conditions that are likely to occur.

Powell and Himelrick (2000) reported successful use ofunder-tree sprinkling with microsprinklers in Alabama and Louisiana on Satsumamandarin. Their goal was to find a method that would provide full protectionagainst moderate frosts and protection to the trunk and lower branches duringsevere frosts. The partial protection during severe frosts helps damaged treesto recover more rapidly. They reported that two risers per tree (i.e. at 0.75 mand 1.5 m), with an output rate of 90.8 litre h^-1 per sprinkler head,gave the best results.

Low-volume (trickle-drip)irrigation

Low-volume (trickle-drip) irrigation systems are sometimesused for frost protection with varied results. Any benefit from applying watercomes mainly from freezing water on the surface, which releases latent heat.However, if evaporation rates are sufficiently high, it is possible that moreenergy can be lost to vaporize water than is gained by the freezing process.Because of the wide variety of system components and application rates, it isdifficult to generalize about the effectiveness of low-volume systems. Again,the best approach is to test the system during the dormant season and note whathappens under a range of weather conditions. If the water on the ground surfaceis a liquid-ice mixture at 0 °C, then the system is beneficial. However, ifall the water freezes and has a milky white appearance, the system wasinefficient for those conditions. One should be aware that operating alow-volume system under frost conditions might damage the irrigation system iffreezing is severe. Heating the water would reduce the chances of damage andwill provide more protection. However, heating may not becost-effective.

Heated water

Davieset al.(1988) reported that water dropletcooling as they fly through the air is the main mechanism of heat supply toorchards during under-plant sprinkling. They hypothesized that freezing water onthe surface to release the latent heat of fusion provides little sensible heatto air (i.e. it does not raise the air temperature). Because of the lowtrajectory of the under-plant spray, evaporation is reduced relative toover-plant systems and preheating water might provide some benefit for theunder-plant sprinklers. Martsolf (1989) applied water heated to 70 °Cthrough a microsprinkler system to a Florida citrus orchard and found littleeffect on the temperature of leaves that were 3 m from the sprinkler heads.However, he found as much as 4 °C rise in temperature of leaves in densetree canopy directly above the heads. On average, temperatures rises variedbetween 1 °C and 2 °C depending on proximity to the sprinkler heads.However, the efficiency resulting from use of a heat exchanger to heat water andthe resulting uniform distribution of energy within the orchard was muchimproved over using point-source orchard heaters. Also, because the watertemperature is low relative to heater temperatures, inversion strength is lessimportant. Where inexpensive energy is available and/or water is limited, theyrecommend using an economical heating system to warm water to about 50 °C.This will lower the required application rate for growers with inadequate watersupplies.

When water is heated to 50 °C, the energy released bycooling to 0 °C and freezing is 544 kJ kg^-1. However, a 2.0 mmh^-1 application rate of 50 °C water gives the same amount ofenergy as a 2.6 mm h^-1 application rate of 20 °C if all of thewater is cooled and frozen. Because of enhanced sensible heat transfer fromwarmer water droplets to the air, heating the water will raise air temperaturein the crop regardless of the frost conditions. However, for growers withadequate water supply and mild to moderate frost conditions, it is probably morecost-effective to design the sprinkler system with the higher application ratethan to pay the additional costs for a heating system, fuel and labour. However,the use of heated water might be a useful alternative for growers with severefrost problems, a source of low cost energy or a limited water supply. Evans(2000) estimates a cost range from $6180 to $8650 ha^-1 for a heatexchanger to heat water for under-plant sprinklers, which is roughly equivalentto twice the cost of wind machines.

Surface irrigation

One of the most common methods of frost protection is todirectly apply water to the soil using furrow, graded border, or floodirrigation. Jones (1924), the earliest known research on using surface water,found a 1 °C increase in air temperature in a citrus grove irrigated withwater at 23 °C. In this method, water is applied to a field and heat fromthe water is released to the air as it cools. The temperature of the water isimportant because warmer water will release more heat as it cools. Protection isbest on the first night after flooding and it becomes less efficient as the soilbecomes saturated. Water can be applied until there is partial or totalsubmersion of tolerant plants; however, fungal disease and root asphyxiation aresometimes a problem. Generally, the method works best for low-growing tree andvine crops during radiation frosts. In an experiment on tomatoes, unprotectedplants showed complete damage (Rosenberg, Blad and Verma, 1983). Usingover-plant sprinkler irrigation gave better protection than with furrowirrigation, but the damage was minor for both methods.

Flooding

Direct flooding is commonly used for frost protection in manycountries. For example, in Portugal and Spain, growers apply a continuous flowof water to a field that partially or totally submerges the plants (Cunha, 1952;Díaz-Queralto, 1971). In Portugal, it has mostly been used to protectpastures of ryegrass and Castilian grass (Cunha, 1952), but it has beensuccessfully used on a variety of crops in California and other locations in theUSA. Because of the relatively low cost of flood irrigation, the economicbenefits resulting from its use for frost protection are high. The volume ofwater to apply depends on the severity of the frost and the water temperature.Businger (1965) indicates that 4 °C of protection can be achieved with thismethod if irrigation is done prior to a frost event; whereas Georg (1979)reports that direct flooding has given temperature rises near 3 °C in apimento pepper crop on a frost night.

Liquid water is denser when the temperature is about 4 °Cthan at lower temperatures, so water at temperatures less than 4 °C willrise to the surface and hence water freezes from the top down. Once the iceforms on the surface, an air space develops between the liquid water below andthe ice above that insulates against the transfer of heat from below. Then theice-covered surface temperature can fall below 0 °C and lead to coldersurface and air temperatures.

Furrow irrigation

Furrow irrigation is commonly used for frost protection andthe basic concepts are similar to flood irrigation. Both free convection of airwarmed by the water and upward radiation are enhanced by flow of warmer waterdown the furrows. The main direction of the radiation and sensible heat flux isvertical, so the best results are achieved when the furrows are directly underthe plant parts being protected.

For example, furrows are needed on the edges of citrus treerows so that air warmed by the furrow water transfers upwards along the treeedges rather than under the trees where the air is already warmer or in themiddles between rows where the air rises without intercepting the trees. Fordeciduous trees, the water should run under the canopy, where the warmed airwill transfer upwards to warm buds, flowers, fruit or nuts. The furrows shouldbe wide so that they present a greater surface area of water. The energy emittedis in W m^-2, so increasing the width of the furrows provides a largersurface area to radiate energy and to heat the air.

Furrow irrigation should be started early enough so that thewater reaches the end of the field before the air temperature falls to thecritical damage temperature. Water at 20 °C will radiate 419 Wm^-2 of energy, whereas water or ice at 0 °C radiates 316 Wm^-2 of energy. Also, the warmer water will transfer more heat to thenearby air, which will transfer vertically into the plant canopy. Ice formationon the water surface insulates the transfer of heat from the water and reducesprotection. With a higher flow rate, the ice formation will occur further downthe row (Figure 7.14), so high application rates afford better protection. Coldrunoff water should not be recirculated. Heating the water is definitelybeneficial for protection; however, heating may or may not be cost-effective. Itdepends on capital, energy and labour costs compared with the potential cropvalue.

FIGURE 7.14
Upward long-wave radiation (Wm^-2) from furrow-irrigation water while it cools and freezes as itflows down a field during a radiation frost night. In the upper figure, thewater cools more rapidly and ice forms closer to the inlet

Foam insulation

Application of foam insulation to low growing crops for frostprotection was widely studied mostly in North America and it has been shown toincrease the minimum temperature by as much as 12 °C (Braud, Chesness andHawthorne, 1968). However, the method has not been widely adopted by growersbecause of the cost of materials and labour, as well as problems with coveringlarge areas in short times due to inaccuracy of frost forecasts (Bartholic,1979). Foam is made from a variety of materials, but it is mostly air, whichprovides the insulation properties. When applied, the foam prevents radiationlosses from the plants and traps energy conducted upwards from the soil.Protection is best on the first night and it decreases with time because thefoam also blocks energy from warming the plants and soil during the day and itbreaks down over time. Mixing air and liquid materials in the right proportionto create many small bubbles is the secret to generating foam with low thermalconductivity. Several methods to produce foam and apply it are reported inBartholic (1979). However, Bartholic (1979) notes that growers show an interestafter experiencing frost damage, but they rarely adopt the use of foam in thelong term. Recently, Krasovitskiet al.(1999) have reported on thermalproperties of foams and methods of application.

Foggers

Natural fog is known to provide protection against freezing,so artificial fogs have also been studied as possible methods against frostdamage. Fog lines (Figure 7.15) that use high-pressure lines and special nozzlesto make small (i.e. 10 to 20 mm diameter) fogdroplets have been reported to provide good protection under calm windconditions (Mee and Bartholic, 1979). Protection comes mainly from the waterdroplets absorbing long-wave radiation from the surface and re-emitting downwardlong-wave radiation at the water droplet temperature, which is considerablyhigher than apparent clear sky temperature. The water droplets should havediameters about 8 mm to optimize the absorption ofradiation and to prevent the water droplets from dropping to the ground. Afairly dense cloud of thick fog that completely covers the crop is necessary forprotection. This depends on the presence of light wind and relatively highhumidity. For example, Brewer, Burns and Opitz (1974) and Itier, Huber and Brun(1987) found difficulties with the production of sufficient water droplets andwith wind drift. Mee and Bartholic (1979) reported that the Mee foggers haveenergy use requirements that are less than 1 percent of heaters, about 10percent of wind machines and about 20 percent of sprinklers. They also reportedbetter protection under some conditions than from use of heaters.

The capital cost for line fogger systems is high, but theoperational costs are low. However, based on personal communications withgrowers and researchers who have tested line foggers in locations with moderateto severe frosts, the fog prevented trees from being killed but it did not savethe crop. Therefore, line foggers should only be used for protection againstmild frost events. In addition, fog drift can be hazardous, so foggers shouldnot be used in locations where car traffic is present.

FIGURE 7.15
An artificial line fogger system operatingin a California almond orchard

Natural fogs that were created by vaporizing water with jetengines have been observed to provide protection. Fog created by the Gillsaturated vapour (SV) gun (Figure 7.16) is considered a natural rather than anartificial fog. The SV gun adds water vapour to the air until it becomessaturated and causes fog to form. The jet engine approach has the advantage thatit can be moved to the upwind side of the crop to be protected. Therefore, thecapital cost of a SV gun is considerably less than for a line fogger system.However, because it has a jet engine, noise is a serious problem. Also, the sameproblem with fog drift exists, so the SV gun should not be used where there iscar traffic. Operation of the machine is somewhat complicated and results fromfield trials have been mixed.

FIGURE 7.16
A Gill saturation vapour gun for natural foggeneration

Combination methods

Wind machines and under-plantsprinklers

Under-plant sprinklers with low trajectory angles can be usedin conjunction with wind machines for frost protection. In addition to heatsupplied by the water droplets as they fly from the sprinkler heads to theground, freezing water on the ground releases latent heat and warms air near thesurface. While this warmed air will naturally transfer throughout the crop,operating wind machines with the sprinklers will enhance heat and water vapourtransfer within the mixed layer to the air and plants. Typically, growers startthe lower cost sprinklers first and then turn on the wind machines if moreprotection is needed. Unlike using heaters with wind machines, the sprinklerheads near the wind machine can be left operating. Evans (2000) reports that thecombined use of wind machines and water can double the benefit of using eithermethod alone. Also, he notes that the combination of method reduces the waterrequirement. Because operating wind machines artificially increases the windspeed, evaporation rates are higher. Consequently, the combination of windmachines and over-plant sprinklers is likely to be detrimental for frostprotection and should not be used.

Wind machines and surfaceirrigation

The combination of wind machines and surface irrigation iswidely practiced in California and other locations in North America, especiallyin citrus orchards. Growers typically start with the surface water and turn onthe wind machines later to supplement protection when needed. As withunder-plant sprinklers, the wind machines facilitate the transfer to the air andtrees of heat and water vapour released from the water within the mixed layer.It is well known by growers that the combination of wind machines and surfacewater improves frost protection. However, the additional amount of protectionafforded is unknown.

Wind machines andheaters

The combination of wind machines and heaters is known toimprove frost protection over either of the methods alone (Martsolf, 1979a). Infact, Brooks (1960) reported that a wind machine and 50 heaters per hectare wasroughly equal to 133 heaters per hectare alone. In California, the combinationof methods was found to be 53 percent, 39 percent and 0 percent cheaper in yearswith 100, 50 and 10 hours of protection, respectively. In California, thecombination has protected citrus orchards to temperatures as low as -5 °Cand only half as many heaters are needed when the two methods are combined. Atypical system has a 74.5 kW wind machine with about 37 evenly spaced stackheaters per hectare, with no heaters within 30 m of the wind machine (Angus,1962). Many efforts to use wind machines to distribute supplemental heat throughor nearby the fans have failed. Fossil fuel heaters placed too close to the fanscause buoyant lifting and decrease wind machine effectiveness. Because the fanoperation tends to draw in cold air near the ground on the outside edge of theprotected area, placing heaters on the outside edge warms the influx of coldair. Placing about half as many heaters (25 to 50 ha^-1) with eachburning oil at a rate of 2.8 litre h^-1on the periphery of the areaprotected by a wind machine saves as much 90 percent of the heater fuel over theseason and improves frost protection because the heaters are not used on manymild frost nights (Evans, 2000). The heaters can be spaced between every secondtree on the outside edge of the orchard and widely spaced within the areaaffected by each wind machine. The concentration should be a little higher onthe upwind side of the orchard. No heaters are needed within about 50 m of thewind machine and the wind machines are started first. If the temperaturecontinues to fall, the heaters are then lit.

Sprinklers and heaters

Although no research literature was found on the use ofsprinklers and heaters in combination, Martsolf (1979b) reported successful useof the combination by a grower in Pennsylvania, USA. The grower had designed acover (i.e. a round metal snow sled mounted horizontally on a pole at about 1.5m above the heater) to prevent water from extinguishing the heater. The growerwould start the heaters first and would only start the sprinklers if the airtemperature fell too low. This combination reduced ice accumulation on theplants and sometimes the sprinklers were not needed. Whether water hitting theheater caused a reduction in heat generation or if it enhanced vaporization andbeneficial fog formation was unknown.