Median Drop Size Research Articles

Dripping Rainfall Simulators for Soil Research—Performance Review

Rainfall simulators represent often-used equipment for soil research. Depending on their performance, they could be appropriate for some soil research or not. The aim of this research is to provide insight into the capabilities of existing dripping rainfall simulators (DRS) to mimic natural rainfall and the frequency of simulated rainfalls of certain characteristics, facilitate the selection of rain simulators that would best meet the needs of soil research and to reach a step closer to the standardization of rainfall simulators. DRS performance was analyzed integrally, for simulators with more than one dripper (DRS>1) and with one dripper (DRS=1). A statistical analysis was performed for the performance of the DRS, wetted area, drop size, rainfall intensity, duration and kinetic energy. The analysis showed that DRS can provide rainfall that corresponds to natural rainfall, except in terms of the drop size distribution and wetted area. However, usually there are more factors that do not correspond to natural rainfall, such as the median drop size, volume and kinetic energy. Metal and plastic tubes (MT and PT) as the most present dripper types showed a strong relation between the outer diameter (OD) and drop size, while the inner diameter (ID) relation was moderate-to-weak. However, when increasing the range of MT drippers, for diameter size, the relation significance becomes very strong for bouts ID and OD. With the increase in the ID of PT, the relation deviates from the logarithmic curve that represents all drippers together. The sizes of the drops generated by the drippers are mostly in the range between 2 and 6 mm, while the number of drops smaller than 2 mm is relatively small. The intensity and duration of the simulated rain can be successfully produced to match natural values, with the most frequently simulated short-term rainfall of a high intensity. Most simulations were conducted at a fall height of up to 2 m, and then their number gradually decreases as the height gets closer to 5 m. Most simulations (58.6%) occur in the range between 20-90% KE, then 33.0% in a range of 90-100%, with only 8.4% lower than 20% KE.

Reducing The Soil Erosion by Improving Infiltration Rate of Sprinkler Irrigation Under Laboratory Conditions

The soil under arid and semi-arid conditions is subject to erosion by irrigation. Polyacrylamide (PAM) significantly reduces erosion from irrigation. The PAM applied with sprinklers will provide erosion control and benefit the water infiltration rate, Soil was sieved by (6.2 mm) and placed in (0.75 * 0.60 * 0.15 m) steel boxes, a slope gradient of (6%), a median drop size of (1.2) mm was sprayed by a hanging oscillating nozzle on a height of (3 m). It used (0, 3, and 6 kg ha-1) PAM sprayed by (80 mmh-1) of water intensity for irrigation, followed by the two other-only irrigations. At (PAM=0) runoff mm, soil loss kgha-1 decreased (40%, 73.15%) respectively in the wheat straw treatments compared with bare flumes, in the time infiltration rate mmh-1 increased by (24.52% )in the same condition, It was shown that a significant difference between the mean of runoff at W. straw and bare treatment at (P <0.05), also a runoff was highly significantly affected (P >0.01) by using (PAM), the result shown that there was a nonlinear relationship between soil loss and runoff. with a coefficient of determination R2=0.95. The costs of (polyacrylamide), are between (40-50 $ ha-1) it depending on the type and the company, W. straw residue can be obtained cheaply after the domestic harvesting season.

Categorization of Sprays by Image Analysis with Convolutional Neuronal Networks

AbstractSpray characterization has been an issue for process and product characterization for decades. Because of this, a convolutional neuronal network was developed to determine the droplet size from spray images. The images were taken using a digital camera, a light source, and a dark room. These were subsequently employed to design and train a convolutional neuronal network using open‐source software packages and a desktop computer. The accuracy of the network droplet size determinations was checked with additional, independent images. The median drop size was assessed with a high accuracy of more than 99.8 % as the mean spray performance indicator. Additionally, the droplet size distribution measurements from the neural network method deviated from those from the reference method (laser diffraction) by less than 1.5 %. Convolutional neuronal networks can be applied to determine the spray performance using spray cone images. This approach could be useful for multiple applications.

Storm-Scale Polarimetric Radar Signatures Associated with Tornado Dissipation in Supercells

Abstract Polarimetric radar data from the WSR-88D network are used to examine the evolution of various polarimetric precursor signatures to tornado dissipation within a sample of 36 supercell storms. These signatures include an increase in bulk hook echo median raindrop size, a decrease in midlevel differential radar reflectivity factor (ZDR) column area, a decrease in the magnitude of the ZDR arc, an increase in the area of low-level large hail, and a decrease in the orientation angle of the vector separating low-level ZDR and specific differential phase (KDP) maxima. Only supercells that produced “long-duration” tornadoes (with at least four consecutive volumes of WSR-88D data) are investigated, so that signatures can be sufficiently tracked in time, and novel algorithms are used to isolate each storm-scale process. During the time leading up to tornado dissipation, we find that hook echo median drop size (D0) and median ZDR remain relatively constant, but hook echo median KDP and estimated number concentration (NT) increase. The ZDR arc maximum magnitude and ZDR–KDP separation orientation angles are observed to decrease in most dissipation cases. Neither the area of large hail nor the ZDR column area exhibit strong signals leading up to tornado dissipation. Finally, combinations of storm-scale behaviors and TVS behaviors occur most frequently just prior to tornado dissipation, but also are common 15–20 min prior to dissipation. The results from this study provide evidence that nowcasting tornado dissipation using dual-polarization radar may be possible when combined with TVS monitoring, subject to important caveats.

Throughfall isotopic composition in relation to drop size at the intra-event scale in a Mediterranean Scots pine stand

Abstract. The major fraction of water reaching the forest floor is throughfall, which consists of free throughfall, splash throughfall and canopy drip. Research has shown that forest canopies modify the isotopic composition of throughfall by means of evaporation, isotopic exchange, canopy selection and mixing of rainfall waters. However, the effects of these factors in relation to throughfall isotopic composition and the throughfall drop size reaching the soil surface are unclear. Based on research in a mountainous Scots pine stand in northeastern Spain, this study sought to fill this knowledge gap by examining the isotopic composition of throughfall in relation to throughfall drop size. In the experimental stand, throughfall consisted on average of 65 % canopy drip, 19 % free throughfall and 16 % splash throughfall. The dynamics of the isotopic composition of throughfall and rainfall showed complex behaviour throughout events. The isotopic shift showed no direct relationship with meteorological variables, number of drops, drop velocities, throughfall and rainfall amount, or raindrop kinetic energy. However, the experiment did reveal that the isotopic shift was higher at the beginning of an event, decreasing as cumulative rainfall increased, and that it also increased when the median volume drop size of throughfall (D50_TF) approached or was lower than the median volume drop size of rainfall (D50_RF). This finding indicates that the major contribution of splash throughfall at the initial phase of rain events matched the highest vapour pressure deficit (VPD) and, at the same time, corresponded to higher isotopic enrichment, which implies that splash droplet evaporation occurred. Future applications of our approach will improve understanding of how throughfall isotopic composition may vary with drop type and size during rainfall events across a range of forest types.

Improvements to the snow melting process in a partially double moment microphysics parameterization

AbstractPolarimetric upgrades to the U.S. radar network have allowed new insight into the precipitation processes of tropical cyclones. Previous work by the authors compared the reflectivity at horizontal polarization and differential reflectivity observations from two hurricanes to simulated radar observations from the WRF model, and found that the aerosol‐aware Thompson‐Eidhammer microphysical scheme performed the best of several commonly used bulk microphysical parameterizations. Here we expand our investigation of the Thompson‐Eidhammer scheme, and find that though it provided the most accurate forecast in terms of wind speed and simulated radar signatures, the scheme produces areas in which the differential reflectivity was much higher than observed. We conclude that the Thompson‐Eidhammer scheme produces drop size distributions that have a larger median drop size than observed in regions of light stratiform precipitation. Examination of the vertical structure of simulated differential reflectivity indicates that the source of the discrepancy between the model and radar observations likely originates within the melting layer. The treatment of number production of rain drops from melting snow in the microphysical scheme is shown to be the ultimate source of the enhancement of differential reflectivity. A modification to the scheme is shown to result in better fidelity of the radar variables with the observations without degrading the short‐term intensity forecast. Additional tests with an idealized squall line simulation are consistent with the hurricane results, suggesting the modification is generally applicable. The modifications to the Thompson‐Eidhammer scheme shown here have been incorporated into updates of the WRF model starting with version 3.8.1.

Erosion Potential under Miconia calvescens Stands on the Island of Hawai‘i

AbstractThis study provides evidence that Miconia calvescens has the potential to accelerate surface erosion in stands where it invades by (i) reducing under‐canopy light levels, thereby reducing the establishment of ground cover vegetation, and (ii) producing highly erosive throughfall drops on large leaves in a single‐layer canopy. The throughfall energy in a stand of invasive miconia on the Island of Hawai‘i (USA), assessed by measuring the drop size and drop velocity distributions with a laser disdrometer, was significantly higher than that in a stand of native ‘ōhi‘a (Metrosideros polymorpha) and ambient rainfall. Median throughfall drop size for miconia (3·83 mm) was twice that of ambient rainfall (1·62 mm). Highly erosive throughfall resulted from large drops forming on large miconia leaves and relatively high fall velocities associated with the single‐story miconia canopy. In contrast, multi‐storied natural ‘ōhi‘a had a larger median drop size; however, a lower fall height reduced throughfall effective kinetic energy. Furthermore, the effective kinetic energy for miconia was high because large drops (&gt; 3·8 mm) with high kinetic energy accounted for 60 per cent of the total energy (versus 30–40 per cent for other vegetation types). Consequently, unit kinetic energy of throughfall was 28 J m−2 mm−1 under miconia, compared with &lt;24 J m−2 mm−1 for rainfall and &lt;20 J m−2 mm−1 under ‘ōhi‘a. These data, combined with the observation of limited protective ground cover under miconia, show the potential for accelerated erosion occurring on forest floors in stands of invasive miconia. Copyright © 2013 John Wiley &amp; Sons, Ltd.

Relationships between the Raindrop Size Distribution and Properties of the Environment and Clouds Inferred from TRMM

Abstract Raindrop size distribution (DSD) retrievals from two years of data gathered by the Tropical Rainfall Measuring Mission (TRMM) satellite and processed with a combined radar–radiometer algorithm over the oceans equatorward of 35° are examined for relationships with variables describing properties of the vertical precipitation profile, mesoscale organization, and background environment. In general, higher freezing levels and relative humidities (tropical environments) are associated with smaller reflectivity-normalized median drop size (εDSD) than in the extratropics. Within the tropics, the smallest εDSD values are found in large, shallow convective systems where warm rain formation processes are thought to be predominant, whereas larger sizes are found in the stratiform regions of organized deep convection. In the extratropics, the largest εDSD values are found in the scattered convection that occurs when cold, dry continental air moves over the much warmer ocean after the passage of a cold front. These relationships are formally attributed to variables describing the large-scale environment, mesoscale organization, and profile characteristics via principal component (PC) analysis. The leading three PCs account for 23% of the variance in εDSD at the individual profile level and 45% of the variance in 1°-gridded mean values. The geographical distribution of εDSD is consistent with many of the observed regional reflectivity–rainfall (Z–R) relationships found in the literature as well as discrepancies between the TRMM radar-only and radiometer-only precipitation products. In particular, midlatitude and tropical regions near land tend to have larger drops for a given reflectivity, whereas the smallest drops are found in the eastern Pacific Ocean intertropical convergence zone.

Microphysics of Raindrop Size Spectra: Tropical Continental and Maritime Storms

Abstract This work uses raindrop size spectra measured at the surface in tropical continental storms to determine the associated parameters of the best-fit gamma distributions. The physical processes responsible for those parameters and their relations to the measurable radar reflectivity Z and differential reflectivity ZDR are then explored. So too are their relations to quantitative measurements of rain. Comparison is then made with corresponding features previously reported in tropical maritime regimes. The storms observed in Brazil and Arecibo, Puerto Rico, have been divided into convective (C), transition (T), and stratiform (S) segments. The raindrop size distribution (DSD) parameters are clearly defined on a gamma parameter diagram (GPD) that shows 1) how median volume drop size D0 increases from S to T to C segments of the rain while 2) the range of the spectrum breadth parameter μ increases, and the range of the slope parameter Λ decreases in the same sequence of S to C. Drop growth occurs predominantly below the 0°C level by collision, coalescence, and breakup in the C rains. The median volume diameter D0 grows as more of the water is concentrated near that size and so the DSD narrows; that is, both μ and Λ increase. In both maritime and continental storms the DSD in the convective portion of the storm approaches equilibrium. The coefficient A in the Z = ARb relation increases with D0 while the exponent b approaches unity. The D0 and A pair increase with, and appear to be determined largely by, the updraft strength, thus providing a possible means of determining the appropriate algorithms for rainfall measurement. Although the small drop number samples measured by the surface disdrometer relative to the large volumes sampled by a radar tend to truncate the DSD at both small and large drop sizes, narrow distributions with μ = 5 to 12 cannot be attributed to such an effect. Such narrow DSDs accord with common experience of monodispersed large drops at the beginning of a convective storm. There is also remarkable agreement of the surface-based observations of ZDR–Z–D0 with the time–space variations from C to T to S rain types observed by radar in England and elsewhere. Because the C region of a storm often accounts for a major share of the rain accumulation despite its shorter duration, it is particularly important to measure that region more accurately. There are distinctive clusters of the generalized number parameter NW versus D0 between maritime and continental storms. Methods for remote sensing and parameterization must partition the rainstorms into convective, transition, and stratiform segments.

Effect of Surfactant Concentration and Operating Temperature on the Drop Size Distribution of Silicon Oil Water Dispersion

The effect of sodium lauryl sulphate (SLES) surfactant and the operating temperature on the drop size distribution of a 350 cSt Dow Corning 200 series oil water dispersion was successfully studied. The dispersion was prepared in a standard 6 litres mixing tank at different impeller speeds. A measurement of the SLES critical micelle concentration (CMC) at 25°C was carried out. The interfacial tension of silicon oil water under various SLES concentration at a temperature range of 25 to 80°C was accomplished. Results showed that the interfacial tension of the silicon oil water decreased as the operating temperature increased and as the surfactant concentration increased. When the operating temperature was increased at the highest SLES concentration tested, a decrease of d 32 was observed. This was attributed to the possibility of hydration of the surfactant at high temperature. Same behavior was observed when measuring the drop size distribution at constant temperature but different SLES concentration. It was found that the mean drop size decreases with mixing time. Different slopes of the change of the median drop size with time were obtained for different SLES concentration. For the same concentration, the slope changes after 1 hour. The degree change of the slope is due to the change of interfacial area of the oil water as mixing time elapsed and the depletion of the surfactant concentration.

High strain rate fragmentation of liquid systems at atmospheric pressure

The fragmentation characteristics of liquid systems at atmospheric pressure has been investigated experimentally and compared to hydrodynamic calculations as well as theoretical predictions. The geometry is a one-dimensional (1-D) nylon flat plate impacting a flat plate liquid system at velocities of approximately 0.3 km/s. The experiments were conducted at the Marquette University's gas gun facility. Hydrocodes calculations were used to investigate early time shock evolution, material deformation and strain rate. High-speed photography and witness cards were used to capture the impact and fragmentation event as well as drop distributions. The experimental drop distributions are compared to distributions obtained from a Grady–Kipp (GK) fragmentation model. The liquid spall GK fragmentation model predicted the correct median drop size while the flow stress spall over predicted the median drop size. The standard GK Poisson drop distribution superimposed over the median drop size produces a distribution which is wider than the data obtained.

Listening to Raindrops from Underwater: An Acoustic Disdrometer

Abstract Different sized raindrops splashing on a water surface produce sound underwater that is distinctive and can be used to measure the drop size distribution in the rain. Five acoustically significant raindrop sizes are described. An inversion of the underwater sound to measure the drop size distribution in the rain is described and demonstrated. Limitations to the inversion include problems associated with the relative loudness of the largest drops (diameter over 3.5 mm), the relative quietness of the medium drops (diameter 1.2–2.0 mm), and the influence of wind to suppress the signal from the otherwise remarkably loud small drops (diameter 0.8–1.2 mm). Various measures of rainfall, including rainfall rate, equivalent radar reflectivity, median drop size, and other integrated moments of the drop size distribution are measured acoustically and used to examine rainfall research issues. The relationship between equivalent reflectivity and rainfall rate, the Z–R diagram, is partitioned acoustically show...

Measurement of simulated drop size distribution with an optical spectro pluviometer: sample size considerations

Knowledge of drop size distributions is important for deriving various rain erosivity parameters. This study investigates the potential of an optical spectro pluviometer (OSP) to measure drop size distributions. Particular attention is paid to the impact of drop sample size and derived erosivity parameters. An experimental set-up using a rainfall simulator and an OSP is described. The OSP allows a continuous real-time sampling of the drops. Results on drop size distributions and sampling effects are discussed. A simulation aimed at reproducing the sampling made with the widely used flour-pellet or filter-paper method is described. From this simulation, recommendations on the sample size of the collected drops needed for an accurate determination of median drop size and kinetic energy are given. Past studies reporting drop size characteristics have often used too small a sample for an adequate description of rain erosivity. Copyright © 1999 John Wiley & Sons, Ltd.

Sprinkler Irrigation Runoff and Erosion Control with Polyacrylamide — Laboratory Tests

AbstractMany semiarid and arid soils are prone to irrigation‐induced erosion. Polyacrylamide (PAM) greatly reduces erosion from furrow irrigation. We hypothesized that PAM applied via sprinklers will provide erosion control and benefit water infiltration and aggregate stability. Screened (6.4 mm) Rad silt loam (coarse silty, mixed, superactive mesic Durinodic Xeric Haplocambid) was placed in 1.5 by 1.2 by 0.2 m steel boxes with 2.4% slope. An oscillating nozzle, 3 m above the soil, produced a median drop size of 1.2 mm diameter. We applied 0, 1, 2, 4, and 6 kg ha‐1 PAM in 20 mm of water in the first irrigation, followed by two 20‐mm water‐only irrigations. In a second test, we applied 0, 2, and 4 kg ha‐1 PAM in 8 mm of water in the first irrigation, followed by two 20‐mm water‐only irrigations. Two kilograms per hectare PAM in the first 20‐mm irrigation reduced runoff 70% and soil loss 75% compared to control. Polyacrylamide in 8 mm of water was less effective. Polyacrylamide in the 20‐mm irrigation did not affect tension infiltration; PAM in the 8‐mm irrigation doubled tension infiltration following the third irrigation. Wet aggregate stability following the first irrigation was greater in all PAM treatments than on the check. With 2 kg ha‐1 PAM in the 20‐mm irrigation, it was 55%; in 8 mm, 77%. Polyacrylamide applied in the first irrigation at low rates effectively reduced runoff and erosion. Erosion was more effectively controlled than runoff.

Drop size distribution and energy load of rain storms at Ibadan, western Nigeria

Rainfall erosivity factors were measured during 1980 and 1981 at Ibadan, Nigeria for two rainy seasons, Season 1 from late March to late July and Season 2 from late August to early November. Rainfall characteristics monitored included drop size distribution, kinetic energy, and momentum using a DISTROMET. Additional factors/characteristics monitored were rainfall amount and intensity using a daily recording rain gauge, and sand splash using an Ellison Cup technique. Rainstorms received in March, June and September had a relatively bigger median drop size ( D 50) ranging from 2.75 to 3.00 mm. Similarly rains in May and September had high kinetic energy of 2000–2500 J/m 2, and high momentum of 300–450 kg/m/s. Kinetic energy of rainstorms was related to D 50 ( r=0.63), 5 min maximum intensity ( r=0.79), 30 min maximum intensity, I 30 ( r=0.83), and rainfall amount ( r=0.98). Therefore, rainfall erosivity may be estimated using data of rainfall amount from the daily recording rain gauge. Regression equations proposed for computing kinetic energy (J/m 2) from rainfall amount ( A in mm) differ among seasons, and need to be validated for different ecoregions of west Africa. The proposed equation is KE=25.80 A+54.58 for Season 1 and KE=29.86 A−43.18 for Season 2.

DROP SIZE MODIFICATION IN BLACK LIQUOR SPRAYS FROM COMMERCIAL NOZZLES USING PULSED FLOW

Improved operation of kraft recovery boilers is expected to result from knowing how to obtain a controlled, well-defined droplet size distribution issuing from the black liquor spray nozzle. Black liquor drop size data were obtained from an environmentally acceptable spray facility capable of delivering liquor at normal firing temperatures using commercial nozzles. Image analysis techniques were developed from high-speed video images which gave good two-dimensional representations of black liquor sprays. Previous work showed that black liquor sprays have a characteristic size distribution which is determined by the fluid mechanical forces breaking up the liquid sheet issuing from the nozzle. This work has centered on applying flow pulsation as an independently controlled force on the sheet breakup process in an attempt to change the drop size distribution. Experiments are described which feature controlled frequency flow pulsations. Results are presented which show the dependence of median drop size on pulsation frequencies up to 450 Hz.

Rainfall erosivity in south-central Nigeria

Rainfall data collected from 1986 to 1990 at Okomu (6° 25′N; 5° 12′E; 76 m above mean sea level), a humid forest area near Benin City, southern Nigeria, were used to evaluate erosivity characteristics. The average annual rainfall was 2048 mm and the distribution pattern was unimodal. 49% of rainfall amount fell at 7.5-minute intensities exceeding 25 mm/h. A maximum 7.5-minute intensity of 240 mm/h was observed at the site. The median rain drop size ( D 50) was 2.3 mm. The mean monthly kinetic energy, using the Wischmeier and Smith (1978) recommended procedure, ranged from 1.5 to 87 MJ/ha; whereas the values ranged from 1.5 to 140 MJ/ha using the Kowal and Kassam (1976) equation. These results suggest that equations developed in the tropics would estimate kinetic energy higher than those developed in the temperate regions. The annual value of erosivity was 18510 MJ·mm/ha·h by the EI 30 index, 216 MJ/ha by the KE ≥ 25 mm/h index and 1329 cm 2/h by the AI m index. The study indicates that rainfall of small and large amounts are capable of causing soil erosion at the site because they often fall at erosive intensities and contain big drops. The complementary role of the small- and large-amount storms made the cumulative erosivity of the rains at the study site very high. The information provided in this study shall be useful in estimating the erosive nature of rain in similar environments and will augment other available information in drawing an iso-erodent map for Nigeria. Furthermore, the erosivity factor for soil loss estimation on monthly basis in the area of study can be quantified by selecting the appropriate index values for soil loss equations.

Zerstäuben von Flüssigkeiten

AbstractAtomizing of liquids. Atomizing of liquids, i.e. the formation of droplets in a continuous gas phase or in vacuum, is accomplished in most cases by atomizer nozzles. Single substance pressure nozzles and pneumatic atomizers or air blast atomizers are alternative devices for that purpose. The physical facts of the partition process of the liquid and two phase flow conditions are now better known thanks to recent investigations. This permits estimation of the effect of the most important parameters on the median drop size by equations of physically based dimensionless groups. The different methods referred to in the literature for determining drop sizes lead to different results and therefore still represent a problem. Nevertheless, an attempt is made to give simplified and physically sound equations for estimating droplet size for the most important kinds of atomizers. They provide the user with an overview of the principal influencing factors and should facilitate choice of a suitable atomizer. Besides the frequently applied methods for drop and spray formation, other processes are also described which are used in special cases.

METHODOLOGY: A SIMPLE RAINFALL SIMULATOR AND TRICKLE SYSTEM FOR HYDRO-GEOMORPHOLOGICAL EXPERIMENTS

A simple rainfall simulator is described which consists of stand-alone sprinkling units with cone jet nozzles that spray downward. With a fall height of 4.57 m and water pressure of 67 kPa, the median drop size of the simulated rainfall is 2.40 mm. Rainfall from two units irrigating a 2 by 2.5 m plot has a total kinetic energy of 0.57 J/m2/s which is approximately 90% of the equivalent energy of natural rainfall at the simulation intensity of 72.4 mm/h. To simulate overland flow, perforated trickle pipes were designed that permit overland flow rates to be readily adjusted and closely controlled by simply varying the input water pressure. In field experiments in southern Arizona, rates between 572 to 1400 cm3/s were generated from a 19-mm pipe. Given their low cost, simplicity, and portability, the rainfall simulator and trickle system should be attractive to researchers working in a wide variety of geomorphic environments.

Drop Size Distributions for Irrigation Spray Nozzles

DROP size distributions for irrigation spray nozzles, such as may be used in low or reduced pressure sprinkler systems, were measured with a calibrated stain technique. Similar data from other sources, measured with photographic or pellet techniques, were also obtained. The distributions were fitted with the upper limit log normal (ULLN) distribution function. ULLN parameters for each distribution are tabulated. Distribution characteristics such as the volume median drop size may be calculated directly from the ULLN parameters. A simple regression model for predicting ULLN parameters as functions of nozzle style, size and pressure is proposed and fitted to data for flooding and smooth flat plate spray nozzles. The fit of model to data was evaluated by comparing measured and predicted values for 50th (median) and 99th (volume) percentile drop sizes, and by directly comparing measured and predicted distribution functions. The distance between functions was defined analogous to the Euclidean distance between points in space, leading to definition of a pseudo r2 for the (functional) regression model. The fit between data and model for the two nozzle types was quite good. The models were used to explore the influence of nozzle size and pressure on drop size distributions for the two types of nozzles.