Abstract

Radiation fog in the Central Valley of California has received very little attention in terms of climatological research. This study uses the Geostationary Operational Environmental Satellite (GOES) nighttime fog product to develop a sequence of images and datasets that reveal patterns of nocturnal radiation-fog development in the Central Valley. Twenty long-lived, spatially extensive radiation-fog episodes, occurring from October through January, were selected for the period of 1997–2000. Mean hourly parameters for fog cover, fog development rate, and vertical development were calculated for the 20 episodes in the Central Valley. The study region is separated into five analysis divisions oriented from south to north for spatial comparisons within the valley. Large-scale radiation fog begins developing before 1800 LST, and rates of development vary widely from south to north. Radiation fog develops earlier and covers a larger area of the southern valley as compared with the central and northern portions of the valley. The horizontal extent of radiation fog is maximized at 0600 LST in the southern valley and near midnight in the central and northern parts of the valley. Vertical development reaches 300 m with regularity in the southern valley. Radiation-fog development of greater than 300 m occurs primarily in the early morning hours. Vertical development “bursts” are also observed in the southern valley during the morning hours. Climatologically important conditions for radiation-fog development in the Central Valley include cool 1600 LST surface temperatures, moisture availability as reflected by 1600 LST dewpoint temperatures, early evening surface cooling trends, the rapidity with which mean relative humidity reaches 90%, and the presence of cool, dry air aloft (700–500 hPa).

Introduction

Radiation fog routinely develops in various parts of the United States throughout the year (Carson and Hardy 1963; Hardwick 1973; Ahrens 1994). These fog episodes have profound implications for human activity and the physical environment. Hazardous ground transportation scenarios resulting from reduced visibility are a major concern in fog-prone areas, and delays in airport schedules are common in areas with frequent radiation fog (Peace 1969; Baker et al. 2002; Swenson 2002). Long-lived spatially extensive radiation-fog episodes also have the potential to deposit suspended and dissolved pollutants, and these large-scale fogs influence the diurnal surface energy budget (McLaren et al. 1988; Oke 1988; Collet et al. 1999). In contrast to potential hazards posed by dense fog, fog water intercepted by vegetation is a major component in the seasonal hydrological balance in regions that frequently experience such episodes (Azevedo and Morgan 1974; Cavelier et al. 1996; Bruijnzeel and Veneklaas 1998).

An area of the United States where the occurrence of long-lived, spatially extensive radiation fog is very common is the Central Valley region of California (Suckling and Mitchell 1988). National Climatic Data Center (NCDC 2002) daily observation records show that, from October through March (for season), individual locations across the Central Valley report upward of 25 fog days per season. These radiation fogs occur primarily with temperatures above 0°C in the Central Valley so that the majority of radiation-fog episodes in the valley are composed of liquid droplets and few are ice fogs. Although radiation fog is the most common winter weather hazard in central California, the phenomenon has been inadequately investigated in a research context. Our limited understanding of the subject is reflected in the small number of published works in the climatological and meteorological literature that focus exclusively on radiation fog in the inland valleys of California (Holets and Swanson 1981).

The purpose of this study is to provide a temporal and spatial framework for understanding radiation-fog development in California. The study will investigate nocturnal formation characteristics of radiation-fog episodes that are both long lived and spatially extensive across the Central Valley. By looking exclusively at nocturnal development of these large episodes, the study also provides a context for using the 4-km spatial resolution Geostationary Operational Environmental Satellite (GOES) nighttime fog product (NFP) for climatological analysis. The GOES NFP is currently in use at National Weather Service field offices and has been cited in the literature primarily when the product is in forecast mode or when the product is undergoing field verification (Eyre et al. 1984; d'Entremont 1986; d'Entremont and Thompson 1987; Ellrod 1994; Wetzell et al. 1996; St. Jean 1997).

Background

Radiation-fog environment

Radiation-fog formation depends on a complex combination of boundary layer and synoptic-scale conditions that often have a diurnal and seasonal nature (Meyer and Lala 1990; Meyer et al. 1986). Fitzjarrald and Lala (1989) describe the fog environment in terms of appropriate boundary layer temperature, humidity, wind speed, and wind direction, as well as near-surface radiative cooling profiles, advection currents, and characteristics of the underlying surface. Radiation-fog initiation is modulated by surface radiative fluxes, vertical mixing of heat and moisture, vegetation interactions, and topographic effects (Duynkerke 1991). Radiation fog may also develop, dissipate, and redevelop during the same nocturnal cycle (Welch and Cox 1986). The process of radiation-fog formation is described for 11 fog episodes in the Chemung River Valley in New York by Pilie et al. (1975b). The authors describe a low-level temperature inversion that develops just after sunset and fluctuates slightly in intensity until fog formation. This inversion is suggested to be most intense in the lowest 0.1 m above the ground surface. The development and timing of the low-level temperature inversion, the rate of fog growth, and valley wind patterns are important factors for fog formation and fog persistence in large mountain–valley climate systems (Moore et al. 1987).

Meteorological processes such as radiative flux divergence and surface dew formation have been modeled to gain insight into the temporal, often nocturnal, aspects of fog onset (Fisher and Caplan 1963; Lala et al. 1975). In addition, microphysical parameters have been used to simulate selected physical interactions in the boundary layer that promote nocturnal fog development (Pilie et al. 1975a; Brown 1980; Musson-Genon 1987). Bott et al. (1990) used a one-dimensional approach to simulate the timing of the fog life cycle and also estimated fog-top height.

Forecasting radiation fog is still somewhat problematic because local topography, moisture availability, vegetation, and soil conditions introduce spatial variability into model results and forecast products (Golding 1993; Meyer and Rao 1999). In addition, the analytical precision that is required to diagnose humidity levels, condensation rates, and radiative exchanges adequately is very demanding in a forecast context (Bergot and Guedalia 1994). A number of approaches, including development of local- and regional-scale fog climatological descriptions, have proven valuable for forecasting onset time and dissipation rates (Court and Gerston 1966; Hardwick 1973; Croft et al. 1997). Numerical forecasting simulations carried out by Guedalia and Bergot (1994) confirm the difficulty in radiation-fog prediction. The simulations emphasized the need for accurate initialization data and very precise nocturnal cooling information.

Satellite remote sensing of fog and low stratus clouds

Once fog has developed, remote observation and analysis of cloud tops and cloud droplet parameters may be made by using satelliteborne sensors and sounders (Arking and Childs 1985; Mead et al. 1989; Menzel and Purdom 1994; Wetzel et al. 1996; Lee et al. 1997). At regional scales, images produced by a number of sensors, including The National Oceanic and Atmospheric Administration GOES and the Advanced Very High Resolution Radiometer (AVHRR) imagers, are adequate to resolve fog development and to convey fog and low stratus boundaries accurately (Gustafson and Wasserman 1976; Bendix 2001).

Introductory meteorology texts routinely display satellite images of radiation fog that extends across the Central Valley of California (Aguado and Burt 2001). These images are used to illustrate both the spatial extensiveness of radiation fog and the utility of spaceborne sensors for viewing such phenomena. Lee (1987) observed Central Valley fog episodes using 4-km-spatial-resolution visible imagery and found variability in fog coverage and fog dissipation characteristics among events and across a single event. Highlighting this variability, Lee (1987) described the “urban clear island effect,” in which fog was absent or dissipated more quickly from urban areas in the Central Valley during widespread fog episodes.

The GOES sensors have the spatial (1 and 4 km) and spectral (0.65, 3.9, and 10.7 μm) resolution to provide both daytime and nighttime observations of cloud-top reflectivity and emission (Rao et al. 1990; Kidder and Vonder Haar 1995). Analysis and forecasting of fog and stratiform-cloud development during daylight hours is practiced using bands in the visible wavelengths (0.65 μm) from the GOES imagers. The 3.9-μm near-infrared band can also be used during daylight hours if the thermal component of the radiance is removed (Arking and Childs 1985; Rawlings and Foot 1990; Kleespies 1995).

To view low stratus clouds and fog at night, two IR bands from the GOES or AVHRR imagers can be used to produce the NFP (Putsay et al. 2001). The GOES NFP calculates the brightness temperature difference between GOES band 4 (10.7 μm) and band 2 (3.9 μm) (Ellrod 1995, 2000). The brightness temperature values for low clouds and fog differ between bands in part because of the emissivity differences induced by cloud thickness and droplet size distribution (Hunt 1973; Pinnick et al. 1978; Arnott et al. 1997). The NFP exploits the lower cloud emissivity in the 3.9-μm band as compared with the 10.7-μm band and calculates a value that not only suggests the presence or absence of fog at single pixel resolution but also estimates the vertical depth of the radiation fog (Ellrod 1994).

Study area

The study area for this project includes the traditional boundaries of the San Joaquin and Sacramento Valleys—together called the Central Valley region (Holets and Swanson 1981; Lee 1987). The Central Valley is relatively uniform in elevation through its center and is bounded by two large topographic barriers—the Sierra Nevada to the east, with elevations that can exceed 4000 m, and the California Coast Range to the west, with elevations to 1500 m (Fig. 1). The Sacramento delta is the largest break in the valley boundary and opens to the west toward the San Francisco Bay. For analysis purposes, the valley is divided into five analysis divisions, each of which contain at least four California Irrigation Management Information System (CIMIS 2000) mesonetwork surface observation stations and a regional population center with a first-order meteorological observation station.

Fig. 1.

Study area (Central Valley of California) including the five subunits (rectangles) used in the analysis of the GOES NFP data

Fig. 1.

Study area (Central Valley of California) including the five subunits (rectangles) used in the analysis of the GOES NFP data

Analysis procedures

Data acquisition

The data for this study are composed of surface observations and GOES images. The NCDC surface data for stations in the Central Valley were used to identify fog events that appeared to be long lived and spatially extensive during fog seasons from 1997/98 through 2001/02. The surface data include current weather reports and visibility estimates. For each fog episode, GOES imager data were acquired from the Space Science and Engineering Center at the University of Wisconsin—Madison. Additional surface meteorological data were acquired from CIMIS. GOES images from 1800 to 0600 LST represent the nocturnal cycle for purposes of this analysis. Local sunset time does vary from October through March over the latitudes of the study area. However, for ease of analysis, the authors used 1800 LST as local sunset and 0600 LST for local sunrise. It is likely that some variability was introduced into the study by generalizing local sunrise and sunset.

The GOES data for each event were downloaded and archived using the Man Computer Interactive Data Access System (McIDAS; Suomi et al. 1983). Image quality control consisted of manual and digital tests for proper coverage and navigation, as well as inspections for image disruptions. In addition to rejection for failure to meet quality-control standards, images were also removed if extensive cirrus cloud cover was present during the nighttime hours. The primary data-quality and image-usability issues consisted of cirrus interruption and sensor malfunctions (no data). Because the GOES 10.7-μm band is sensitive to dust (GOES dust identification uses 10.7–12-μm product), it is reasonable to assume that on occasion the 10.7-μm channel may have been influenced by the dust that is commonly suspended in the lower troposphere over the Central Valley. For this study, however, dust concentration was not considered when using the 10.7- and 3.9-μm infrared images.

Identification of events

Hourly observations of current weather and visibility conditions from Bakersfield (BFL), Fresno (FAT), Merced (MER), Sacramento (SMF), and Chico (CIC), California were used to identify long-lived, spatially extensive radiation-fog episodes. These stations span the Central Valley from south to north and represent areas of population concentration. The “current weather” reports were examined for these stations for fog seasons from 1997/98 through 2001/02. An extensive fog episode was identified when at least three of the five stations simultaneously reported fog with visibility reduction to 1 mile for three consecutive nocturnal hours.

Nighttime-fog-product calculation

Calculation of the GOES NFP is discussed in detail in Ellrod (1995). In the current study, the GOES NFP is processed using IR images that were calibrated to a standard navigation and projection (Mather 1999). The temporal pairs of 10.7- and 3.9-μm IR images produce the GOES NFP as follows:

 
ΔT = (Tβ10.7Tβ3.9),
(1)

where Tβ10.7 is the brightness temperature for GOES IR channel 4, Tβ3.9 is the brightness temperature for GOES IR channel 2, and ΔT represents the difference in brightness temperature between the two IR images.

Positive ΔT values represent pixels with the GOES NFP fog signature. Table 1 details the vertical-fog-development estimates associated with the GOES NFP ΔT values. Satellite-estimated fog and stratus depth have been compared with aircraft observations in areas outside the Central Valley over the period of 1997–2001 (Ellrod 2000). Results from a linear regression analysis using 73 pairs of satellite estimates and aircraft observations produced a statistically significant r2 value of 0.625 for the relationship (ΔT vs cloud depth at 100-m intervals) as listed in Table 1 (Ellrod 2000).

Table 1.

Estimates of radiation-fog depth associated with particular ΔT values. The estimates are based on the work of Ellrod (1994, 2000) and have not been field verified in the Central Valley of California

Estimates of radiation-fog depth associated with particular ΔT values. The estimates are based on the work of Ellrod (1994, 2000) and have not been field verified in the Central Valley of California
Estimates of radiation-fog depth associated with particular ΔT values. The estimates are based on the work of Ellrod (1994, 2000) and have not been field verified in the Central Valley of California

Analysis of radiation-fog characteristics

A brief synoptic analysis was performed to determine the general circulation patterns that exist during the nocturnal fog-development cycle. The analysis was limited to surface conditions, 850-hPa heights and wind, and 500-hPa heights and wind for the eastern Pacific Ocean and the western United States for the dates of spatially extensive fog episodes (Yarnal 1994).

To illustrate the spatial and temporal character of radiation fog, the GOES NFP images for a representative fog event were sequenced from sunset to sunrise and were displayed with fog-depth contours. The spatial extent of each fog episode was measured using the <100-m contour. The major and semimajor axes of continuous fog coverage were identified from the edges of the <100-m contour.

To describe further the spatial character of radiation fog in the Central Valley, the study area was partitioned into five analysis divisions whose dimensions are 35 × 35 pixels (GOES 4-km pixels). The divisions were arranged south–north. The fog cover percentage Cp was calculated for each division as follows:

 
Cp = ΔTobsTmax,
(2)

where ΔTobs is the observed number of pixels with ΔT values greater than or equal to 1, and ΔTmax is the maximum number of pixels, out of 1225 (35 × 35 pixels), that are covered with GOES NFP–identified fog for a particular analysis division. Because the five analysis divisions do not have equal distributions of foggy pixels, the Cp allows fog cover to be analyzed in each division based on the observed maximum coverage in that division, thus providing uniformity in analysis.

Time-relative mean fog-depth values were also calculated at the pixel scale for each of the five analysis divisions. Mean fog-depth values were generated for the period of 1800–0600 LST. These values were used to describe the coverage characteristics and rates of fog development. The vertical development characteristics were addressed using the vertical development ratio, which estimates the proportion of pixels covered with “deeper” fog as compared with pixels covered with more “shallow” fog layers in an analysis division. The vertical development ratio (ΔTR) takes the form of

 
ΔTR = (ΔT4 + ΔT5 + ΔT6)/(ΔT1 + ΔT2 + ΔT3),
(3)

where ΔT1, ΔT2, and ΔT3 represent the frequency of pixels with a GOES NFP fog signature for shallow radiation fog, and ΔT4, ΔT5, and ΔT6 represent the frequency of pixels with the GOES NFP signature for deeper-developing radiation fog. A ratio value very close to 0 represents an analysis in which shallow fog dominates the study area. A ratio of 1.0 represents a situation in which shallow fog and deep fog cover an equal portion of the study area, and ratio values greater than 1.0 represent scenarios in which deeper radiation fog covers more of the analysis division.

Last, mesoscale surface meteorological observations were used to determine the nocturnal characteristics of selected variables during extensive fog development. The variables used in this portion of the analysis were obtained from the CIMIS network. The meteorological variables are surface temperature (°C), dewpoint temperature (°C), dewpoint depression (°C), wind direction (degrees from north), and wind speed (m s−1). Hourly observations from four CIMIS stations per analysis division were used to produce mean fog-episode meteorological values for each division. The trends in the mean meteorological values act as surrogates for physical processes that occur in the boundary layer. In the discussion that follows, the meteorological observations are used to suggest the processes that lead to fog onset and maintenance across the study area.

Findings

Overview of the 20 radiation-fog episodes

Identification of long-lived spatially extensive fog episodes, using the criteria set out above, produced 31 events from the fog seasons of 1997–2002. Twenty of the 31 events met the requirements for spatial extent and image quality and had cirrus-free coverage from 1800 through 0600 LST. It is likely that a number of spatially extensive and long-lived episodes have been omitted from this analysis using such restrictive criteria. However, 20 nocturnal fog episodes is a sufficient sample for a detailed radiation-fog analysis, because past radiation fog studies have included as few as one episode (Fitzjarrald and Lala 1989).

Table 2 provides a brief description of each of the 20 radiation-fog events. The mean south–north extent of fog across the study area is 452 km, and the mean east–west extent is 155 km. For the 20 events that make up the current study, the most prominent synoptic-scale surface feature during fog development is a high pressure center usually located near 42.3°N, 117.8°W. The average surface pressure associated with this high pressure center is 1033 hPa, with a range from 1019 to 1045 hPa over the 20 episodes. The mean upper-air pattern, as described by the 500-hPa flow, consists of a highly amplified ridge with its axis located near 120.8°W. Geopotential heights at the 500-hPa level across the Central Valley are estimated from 5550 to 5850 gpm, with a mean value of 5718 gpm during fog development.

Table 2.

General meteorological characteristics associated with the development of the 20 radiation-fog events analyzed in this study. The spatial extent of fog cover was estimated using GOES NFP images with contoured depth intervals

General meteorological characteristics associated with the development of the 20 radiation-fog events analyzed in this study. The spatial extent of fog cover was estimated using GOES NFP images with contoured depth intervals
General meteorological characteristics associated with the development of the 20 radiation-fog events analyzed in this study. The spatial extent of fog cover was estimated using GOES NFP images with contoured depth intervals

Central Valley radiation-fog case study (25 November 2000)

Figure 2a is the 1800 LST GOES NFP image for 24 November 2000. This frame illustrates the initial stage of nocturnal radiation-fog development (near local sunset). Fog development is spatially limited at this time, with fog west of FAT extending 227 km south–north and 147 km east–west. The fog is vertically developed to approximately 300 m. The contours around the main body of fog represent fog that is less than 100-m thick.

Fig. 2.

The (a) 1800 LST 24 Nov 2000, (b) 0000 LST 25 Nov 2000, and (c) 0600 LST 25 Nov 2000 GOES NFP images of developing radiation fog. The contour lines represent fog depth, and five Central Valley cities that lend their names to the analysis divisions are identified (see text for definitions)

Fig. 2.

The (a) 1800 LST 24 Nov 2000, (b) 0000 LST 25 Nov 2000, and (c) 0600 LST 25 Nov 2000 GOES NFP images of developing radiation fog. The contour lines represent fog depth, and five Central Valley cities that lend their names to the analysis divisions are identified (see text for definitions)

Late-afternoon temperatures on 24 November 2000 at both FAT and MER are lower when compared with the other valley divisions. At FAT 1400 LST temperature is 11.7°C, and at MER it is 9.8°C. The other stations in the study report 1400 LST temperatures of 14.5° (BFL), 15.4° (SMF), and 13.7°C (CIC). The colder afternoon temperatures in the middle portion of the valley (FAT and MER) suggest that these areas should develop radiation fog more rapidly with even marginal radiative cooling. With no sounding data available for the Central Valley proper, the Oakland, California (KOAK), sounding was used to assess the vertical profile of the lower troposphere. The 0000 UTC (1600 LST) and 1200 UTC (0400 LST) soundings show westerly flow from the 850- through 500-hPa level (Fig. 3). The westerly flow aloft is very dry, with 700-hPa dewpoint depression at 10°C at 1600 LST. This dry air aloft was also observed on other western U.S. soundings [Deer Park, Washington (KDEW), and Mercury, Nevada (KDRA)] that are east of the Central Valley. A large high pressure center with closed isobars is located over the Great Basin at sunset on 24 November 2000 as well. The central surface pressure for this Great Basin high is 1030 hPa, and the high pressure dome extends prominently through the 850-hPa level. Wind speeds at 850 hPa were estimated at 2.5 m s−1 across California from 1600 to 0400 LST.

Fig. 3.

KOAK soundings for 1600 LST 24 Nov 2000 and 0400 LST 25 Nov 2000

Fig. 3.

KOAK soundings for 1600 LST 24 Nov 2000 and 0400 LST 25 Nov 2000

Figure 2b is the GOES NFP image for midnight during the 25 November 2000 fog episode. Radiation fog covers a large portion of the San Joaquin Valley at this time. The south–north extent is approximately 319 km, and the east–west coverage is 126 km. Fog has also formed near BFL at midnight, and there are patches of fog extending northward from MER. Rapid radiative cooling from 1800 to 0000 LST at BFL produces surface temperatures of 2.1°C at midnight. This temperature observation is more similar to the 0000 LST temperature of 2.2°C at FAT. Temperatures at both SMF and CIC, however, do not cool substantially from sunset to midnight. The SMF temperature at 0000 LST is 6.1°C, and the CIC temperature is 5.6°C. The slower cooling trends in the northern valley may in part explain the lack of radiation-fog development during the evening hours.

Vertical development at 0000 LST is similar to that exhibited at 1800 LST, with fog tops to 200 m across the central portion of the valley. The fog cover extending southward is vertically developed to less than 100 m, while the fog east of SMF has developed to 100 and 200 m over limited areas. Also of note are the embedded pixels that suggest vertical development to 300 m within the main body of fog from FAT to MER.

Figure 2c illustrates spatially extensive radiation fog across most of the Central Valley near sunrise (0600 LST) on 25 November 2000. The south–north axis of fog cover now measures 630 km, and the east–west span is 175 km near SMF. This particular fog episode extends across the Sacramento delta and into the San Francisco Bay area. The fog in the bay area cannot be assumed to have formed under the same radiative conditions as the valley fog to the east, and so this is treated as a hybrid fog episode west of SMF.

All five of the valley population centers are within the main body of fog at 0600 LST, and fog depth has increased since midnight. Extensive areas of 300-m development are indicated between BFL and FAT, and also north of MER. The extreme northern portions of the Sacramento Valley near CIC have vertical development indicated to 200 m with a smaller area to 300 m. Surface cooling during the morning hours produces temperatures of −0.2°C at BFL, 0.6°C at FAT, 1.4°C at MER, 4.3°C at SMF, and 5.8°C at CIC at 0600 LST. The 1200 UTC (0400 LST) KOAK sounding indicates a strong drying profile from approximately 900 to 550 hPa. The winds through this level are very weak and are from the west-southwest. This profile, along with building high pressure over the Great Basin and weak surface winds, suggests that processes are in place that favor maintenance of the fog cover over the Central Valley.

Nocturnal characteristics of radiation fog—Mean spatial coverage

The area covered by radiation fog in each analysis division is represented by the coverage percentage (Cp) whose calculation is detailed in Eq. (2). The maximum observed coverage, from which Cp is calculated, is 1196 pixels for the BFL division, 1114 for FAT, 1085 for MER, 933 for SMF, and 790 for the CIC division.

Figure 4 shows the Cp calculated for each of the analysis divisions. The BFL division reaches 50% mean coverage at 0300 LST and remains above this level for the remainder of the nocturnal cycle, peaking in coverage at 0600 LST. The standard deviation around the hourly mean Cp at BFL is variable from 1800 to 0000 LST but becomes more stable after midnight.

Fig. 4.

Areal coverage percentages (Cp) for each of the five analysis divisions. The Cp values were calculated based on the maximum observed coverage (pixels with GOES NFP fog signature) for each division

Fig. 4.

Areal coverage percentages (Cp) for each of the five analysis divisions. The Cp values were calculated based on the maximum observed coverage (pixels with GOES NFP fog signature) for each division

Maximum coverage at FAT is realized at 0600 LST at 53%. Lower Cp standard deviations are realized during the morning hours when the areal coverage is greatest, suggesting that coverage variability is minimized as coverage increases and meteorological conditions become more favorable for fog development and maintenance.

The MER division peaks in Cp at 0000 LST and again between 0400 and 0500 LST. The standard deviation values suggest that the morning hours are less variable than the evening hours in terms of radiation-fog coverage.

Mean Cp at SMF peaks at 2300 LST at 35%, and the final 4 h of the study period are nearly equal. Similar to the southern valley analysis, the SMF standard deviation around the hourly mean Cp suggests that fog development during the morning hours is less variable than during the evening hours.

The most extensive coverage at CIC occurs at 2300 LST at 32%. The hourly standard deviation analysis suggests that spatial coverage is variable at all hours, with no preference for evening or morning. This variability in Cp reflects the mean surface meteorological observations at CIC, where temperatures during the 20 fog episodes did not cool as rapidly as they did in the other divisions (Fig. 5). In addition, mean relative humidity was lower than in the other divisions and the mean wind speed was higher, remaining above 2.0 m s−1 over the entire nocturnal cycle. The surface mixing induced by higher wind speeds could offset radiative cooling and keep nocturnal temperatures and dewpoint depression values higher at CIC.

Fig. 5.

Mean hourly surface temperature (°C) trends for each of the analysis divisions. The mean values were calculated using four CIMIS stations in each analysis division

Fig. 5.

Mean hourly surface temperature (°C) trends for each of the analysis divisions. The mean values were calculated using four CIMIS stations in each analysis division

The Cp trends in the divisions other than CIC are most likely attributable to the dry, cool, and clear midtroposphere, which promotes radiative cooling throughout the valley, and greater moisture availability in the southern valley. Figure 6 shows that late-afternoon dewpoint temperatures at both BFL and FAT (southern valley divisions) are much higher than in the other analysis divisions. The mean relative humidity trends are similar across the study area from 1600 through 1900 LST but diverge noticeably from 1900 through 0700 LST (Fig. 7). The BFL and FAT analysis divisions rapidly reach mean relative humidity values of 90% while MER, SMF, and CIC are delayed in reaching 90% relative humidity. Radiation-fog development follows a similar trend as the cooling and humidity trends across the divisions—forming more readily and maintaining coverage throughout the nocturnal cycle across BFL and FAT and developing more slowly and dissipating slightly during the morning hours across MER, SMF, and CIC.

Fig. 6.

Mean hourly surface dewpoint temperature (°C) trends for each of the analysis divisions. The mean values were calculated using four CIMIS stations in each analysis division

Fig. 6.

Mean hourly surface dewpoint temperature (°C) trends for each of the analysis divisions. The mean values were calculated using four CIMIS stations in each analysis division

Fig. 7.

Mean hourly relative humidity (%) trends for each of the analysis divisions. The mean values were calculated using four CIMIS stations in each analysis division

Fig. 7.

Mean hourly relative humidity (%) trends for each of the analysis divisions. The mean values were calculated using four CIMIS stations in each analysis division

Nocturnal characteristics of radiation fog—Mean rate of fog development

Figure 8 illustrates the hourly rates of fog-cover change for each of the five analysis divisions. The divisions are represented hourly from top to bottom on the graphic in the following order: BFL, FAT, MER, SMF, and CIC. The hourly rate of coverage change for the BFL division is positive for 10 of the 12 hourly intervals. Radiation fog growth is less intense after midnight, with only 1 h with a rate greater than 5%, as compared with 2 h with rates greater than 7% from 1800 to 0000 LST. The fog expansion rate is also less variable after midnight, reflecting the stability of the environment in terms of air temperature and dewpoint depression (Figs. 5 and 9, respectively). Wind speeds are also at a nocturnal low during the morning hours at BFL (Fig. 10).

Fig. 8.

Radiation-fog development rates for each of the five analysis divisions. The rates were calculated based on hourly changes in Cp. The divisions are represented from top to bottom (BFL, FAT, MER, SMF, CIC) for each hour

Fig. 8.

Radiation-fog development rates for each of the five analysis divisions. The rates were calculated based on hourly changes in Cp. The divisions are represented from top to bottom (BFL, FAT, MER, SMF, CIC) for each hour

Fig. 9.

Mean hourly dewpoint depression (°C) trends for each of the analysis divisions. The mean values were calculated using four CIMIS stations in each analysis division

Fig. 9.

Mean hourly dewpoint depression (°C) trends for each of the analysis divisions. The mean values were calculated using four CIMIS stations in each analysis division

Fig. 10.

Mean hourly surface wind speed (m s−1) trends for each of the analysis divisions. The mean values were calculated using four CIMIS stations in each analysis division

Fig. 10.

Mean hourly surface wind speed (m s−1) trends for each of the analysis divisions. The mean values were calculated using four CIMIS stations in each analysis division

The FAT division has 10 h with positive rates (Fig. 8). The two negative intervals are 0000–0100 and 0100–0200 LST. FAT experiences its greatest fog-cover expansion from 1800 to 1900 LST, which coincides with a rapid temperature decline and increasing relative humidity levels. Even though the calculation of mean values for the 20 episodes dilutes the finer details of each episode's nocturnal cycle, the mean values do suggest that FAT experiences increasing wind speeds between 0000 and 0200 LST (Fig. 10). This period also exhibits a midcycle erosion of radiation fog at FAT, which suggests that some low-level mixing may be present during the early morning. This process ceases, however, and positive fog-development rates recover from 0300 LST to sunrise.

The MER division has 8 of 12 intervals with positive rates of fog development. All four negative rates are during morning intervals. MER is somewhat unique in that the division's mean relative humidity levels remained below 90% until 0300 LST. This observation is reflected in the dewpoint temperature trend, which shows MER having the lowest mean dewpoint temperature for any division from midnight until 0600 LST (Fig. 6).

The SMF division has nine hourly intervals with positive radiation-fog growth rates. The highest growth rate is 7.75% at 2100–2200 LST, and the most negative rate is 5.9%. Temperature and relative humidity trends at SMF are more similar to those in BFL and FAT, and the fog growth rates also follow a trend more similar to the southern valley divisions.

The CIC division has positive radiation-fog growth during 8 of the 12 hourly intervals. The highest positive rate at CIC is 4.4% at 2100–2200 LST. The lowest negative growth is found during 2300–0000 LST at 4.1%. The rates are comparable to the other divisions; however, the total maximum coverage at CIC is still much less than the other divisions (790 pixels).

This analysis suggests that radiation-fog development rates are positive across all divisions during the evening hours and that positive development rates are greatest in the southern divisions (BFL, FAT). Surface meteorological conditions in the southern valley, such as cool 1600 LST surface temperatures and higher moisture levels (dewpoint temperatures), allow the BFL and FAT divisions to develop radiation fog very rapidly after sunset, as revealed by the GOES NFP. Negative rates of fog-cover change are most likely during the middle of the nocturnal cycle. Midcycle wind speed increases at FAT and MER suggest that in the larger east–west expanses of the Central Valley (nearer FAT and MER) low-level mixing is available to erode a portion of the fog cover.

Nocturnal characteristics of radiation fog—Vertical development

Figure 11 is a graphical summary of the mean vertical development ratio ΔTR for each of the five analysis divisions from 1800 to 0600 LST. More shallow fog development dominates the spatial and temporal distribution in the 20 fog episodes; however, variability is evident in the ΔTR from south to north and from sunset to sunrise.

Fig. 11.

Hourly vertical development ratio (ΔTR) for the five analysis divisions. The ΔTR was calculated using pixels with deep developing radiation fog (greater than 300 m) and pixels with more shallow developing radiation fog (less than 300 m)

Fig. 11.

Hourly vertical development ratio (ΔTR) for the five analysis divisions. The ΔTR was calculated using pixels with deep developing radiation fog (greater than 300 m) and pixels with more shallow developing radiation fog (less than 300 m)

The northernmost analysis areas vary little in their ratio of shallow to deeper fogs. Both the SMF and CIC divisions remain at or below a ΔTR value of 0.10 for the entire nocturnal cycle. This result suggests that deep radiation fog rarely develops over large areas of the division. The lack of fog development is likely a result of surface conditions such as lower dewpoint temperatures and higher mean wind speeds. This peak in ΔTR (0.20) is followed by a steep decline in the vertical development ratio. A minimum value of 0.08 is reached at 0200 LST.

The BFL division is dominated by shallow fog cover from 1800 to 2100 LST when the ratio is minimized at 0.03. From 2100 to 0100 LST the ΔTR increases to 0.30 and then experiences another steep increase from 0400 to 0600 LST. The 0600 LST peak ratio is 0.49.

The FAT division also begins the night with ΔTR values near 0.11, but the ratio begins to increase between 1900 and 0000 LST. FAT reaches one of its two peaks at midnight with a ΔTR of 0.30. Just after midnight, the deeper fogs decrease as a proportion of total coverage, but they increase again to a second peak of 0.31 at 0600 LST.

The surface conditions that help to explain the development of deeper fog in the southern portion of the study area have been discussed in some detail in the sections above. They include the south–north variation in 1600 LST temperature, 1600 LST relative humidity levels, the steep temperature decline after sunset, and low mean wind speeds at the surface. It is also very likely that conditions in the midtroposphere play a critical role in vertically developing radiation fog in the Central Valley.

The KOAK soundings for the 20 episodes reveal that the atmosphere above 700 hPa was consistently dry and cooled slightly during radiation-fog development. Cool, dry air above the valley inversion, along with light winds aloft, promotes fog-top radiative cooling, strengthening the temperature inversion, and encouraging further condensation at the fog top. The mean 700-hPa dewpoint depression for the 20 fog episodes was 14.5°C at 1600 LST, and the mean 700-hPa dewpoint depression at 0400 LST was 21.3°C, suggesting significant drying of the air above the fog level in the valley. The mean 500-hPa dewpoint depression value was 12.5°C at 1600 LST and 18.9°C at 0400 LST, confirming that a deep layer of the midtroposphere experienced substantial drying over the nocturnal cycle. Wind speeds at 700 hPa were generally weak (7.7–15.4 m s−1) and were out of the northwest most prominently at both 1600 and 0400 LST.

To summarize the relationship between vertical development parameters and horizontal fog cover, a Pearson's product moment correlation analysis was performed at the 0.05 significance level to compare the hourly vertical development ratio with hourly coverage percentage and hourly fog growth rates. The analysis suggests that there is no significant relationship in terms of linear correlation between vertical fog development and hourly growth rates. This is the case over all five analysis divisions. There are, however, significant correlation coefficients for the relationship between hourly Cp and the hourly ΔTR. For the BFL division, 13 of the 20 fog episodes had significant positive correlations between hourly Cp and the hourly ΔTR. The remaining seven fog episodes had no significant linear relationship at the 0.05 level. The FAT division had 14 fog episodes with positive correlations between the hourly Cp and hourly ΔTR. In the MER division there were 10 fog episodes that had significant correlation coefficients, and each was positive. The SMF division had 11 fog episodes with significant positive correlation coefficients. The CIC division produced only 3 of 20 fog episodes with significant correlations between the hourly Cp and hourly ΔTR. This result suggests that vertical development follows horizontal development more strongly in the southern part of the valley (BFL and FAT), but the relationship weakens in the central part of the study area (MER and SMF) and is nonexistent in the northern valley (CIC).

Summary and conclusions

This study presents a framework for analyzing the nocturnal development of long-lived spatially extensive radiation fog in the Central Valley of California. By dividing the study area into five divisions and utilizing the GOES NFP to analyze 20 fog episodes, the study provides insight into both the spatial and temporal characteristics of large-scale radiation-fog development. This analysis is the first to provide parameters for areal and vertical development of Central Valley radiation fog. The study also links fog-development characteristics to surface and lower-tropospheric conditions during the nocturnal cycle.

Synoptic-scale conditions apparent during the 20 radiation-fog episodes are dominated by a large Great Basin high pressure center. The mean location and central pressure for the omnipresent Great Basin high were 42.3°N, 117.8°W and 1033.6 hPa, respectively. The mean position of the 500-hPa ridge axis associated with this regime was approximately 120.8°W, and 500-hPa heights across the Central Valley during radiation-fog episodes ranged from 5550 to 5850 gpm.

The mean south–north extent of the 20 episodes was found to be 452 km, and the mean east–west extent of radiation fog was 155 km. Even though these large-scale fog events may be perceived to be homogeneous in character, the spatial coverage, rate of development, and vertical extent were found to vary substantially from south to north.

Fog cover as measured by Cp was greatest in the BFL and FAT divisions. Over the 20 episodes analyzed in this study, areal coverage was more variable during the evening hours but became less variable from 0000 to 0600 LST. This was the case for all divisions except CIC. Total radiation fog cover was maximized at 0600 LST in the southern valley but was at its greatest areal extent between 2300 and 0000 LST for MER, SMF, and CIC. The FAT and MER divisions experienced slight erosion of their fog cover between 0000 and 0200 LST. This erosion coincided with an increase in mean wind speed in this area of the valley, which is the widest portion of the study region and, thus, may be subject to circulation patterns influenced by the distance from the center of the valley to adjacent topographic barriers.

Vertical fog development was also variable from south to north across the study area. The deepest fog development occurred in the southern valley, with maximum depths that exceed 400 m. Vertical development was most rapid in the south valley from 2100 to 0100 LST and again between 0400 and 0600 LST. Vertical development was also strongly related (positively) to areal coverage in the southern valley, but this positive relationship was weaker from MER to CIC.

Both the areal extent and the vertical development of radiation fog in the southern valley closely followed the temporal trends in surface temperature, dewpoint temperatures, and relative humidity. Divisions including BFL and FAT also recorded much cooler 1600 LST surface temperatures in comparison with the other divisions. BFL and FAT also had more vigorous radiation fog development in comparison with MER, SMF, and CIC. Vertical development was also influenced by cool, dry air aloft, which was observed during each of the 20 episodes from 700 to 500 hPa. Though the height of the valley temperature inversion was not available from the KOAK sounding, the cool, dry northwesterly winds in the layer above 700 hPa likely strengthened the inversion and modulated condensation at the fog top. The mean dewpoint temperature change from 1600 to 0400 LST at 700 hPa was 6.8°C, and drying at 500 hPa was 6.4°C during the same period.

This work confirms a number of conclusions from radiation fog research carried out in other regions. Among these confirmations are the relationship of nocturnal fog development to presunset temperature conditions, evening trends in relative humidity, and surface wind speed (Fitzjarrald and Lala 1989; Guedalia and Bergot 1994). Further work in the operational context can advance the use of the GOES NFP for both forecasting and climatological analysis in the Central Valley. Foremost among the operational goals should be the verification of the GOES NFP vertical development estimates using visual observations of known topographic features, visibility reports, and GOES or AVHRR visible images.

The findings from this GOES NFP analysis of the Central Valley may have implications for forecasters and public-safety officials in the region. The most promising findings that may make the transition to operational use follow:

  • More accurate timing of areal radiation-fog expansion may prove beneficial, as assumptions of gradual development could lead to spurious forecasts. In the southern part of the valley, using Cp or hourly development rates at midnight to estimate sunrise coverage may underestimate the maximum coverage because rapid development is evident during the morning hours.

  • The identification of a prominent midcycle disruption in fog development in the valley region near FAT and MER may also benefit forecasters. This disruption in fog development signals a retreat in areal coverage over the morning hours near MER and prevents fog development near FAT from reaching the extent of fog episodes farther south near BFL.

  • The rapid “bursts” of vertical fog development after 0100 LST at BFL and after 0200 LST at FAT may allow forecasters in the southern valley to delay analysis of fog depth until later in the diurnal cycle so as not to underestimate sunrise fog depth.

  • The study's confirmation of the importance of analyzing surface temperatures and dewpoint temperatures before sunset (in this case 1600 LST) to estimate time of onset and rapidity in radiation fog development may assist forecasters and planners in preparing aviation forecasts (especially calculating takeoff/landing cessation times at valley airports).

  • In a climatological context (i.e., calculating mean values over many episodes), forecasters may identify the hour at which mean surface relative humidity levels reach 90% and use this hour as an indicator for forecasting rapidly expanding or vertically developing radiation fog.

This study was one of the first to use GOES NFP images for analysis of the climatological aspects of radiation-fog development. The results of this study suggest that the GOES NFP can be a an effective data-generating tool and, in concert with surface and upper-air data, can assist in evaluating both horizontal and vertical development characteristics of radiation fog. The findings should lead to a number of questions that can be approached with similar methods as employed in this study. Among the most intriguing research questions raised by this study is whether the GOES NFP–generated 0600 LST depth estimates can be used to forecast postsunrise time requirements for fog dissipation in the Central Valley. By using the same five analysis divisions and GOES 4- and 1-km visible imagery, this question could be tested with great accuracy.

Acknowledgments

This work was funded by the National Oceanic and Atmospheric Administration's National Environmental Satellite Data and Information Service, Office of Research Applications (NOAA/NESDIS/ORA) under Grant NA06EC0205. Special thanks are given to Kris Lynn (University of California, Kearny Agriculture Center) and the Geography Department at California State University, Fresno.

REFERENCES

REFERENCES
Aguado
,
E.
and
J. E.
Burt
.
2001
.
Understanding Weather and Climate.
Prentice Hall, 505 pp
.
Ahrens
,
D.
1994
.
Meteorology Today. 6th ed.
Brooks-Cole, 535 pp
.
Arking
,
A.
and
J. D.
Childs
.
1985
.
Retrieval of cloud cover parameters from multispectral satellite images.
J. Climate Appl. Meteor.
24
:
322
333
.
Arnott
,
W. P.
,
C.
Schmitt
,
Y.
Liu
, and
J.
Hallett
.
1997
.
Droplet size spectra of water vapor concentration of laboratory water clouds: Inversion of Fourier transform infrared (500–5000 cm−1) optical-depth measurement.
Appl. Opt.
36
:
5205
5216
.
Azevedo
,
J.
and
D. L.
Morgan
.
1974
.
Fog precipitation in coastal California forests.
Ecology
55
:
1135
1141
.
Baker
,
R.
,
J.
Cramer
, and
J.
Peters
.
2002
.
Radiation fog: UPS Airlines conceptual models and forecast methods.
Preprints, 10th Conf. on Aviation, Range, and Aerospace Meteorology, Portland, OR, Amer. Meteor. Soc., 154–159
.
Bendix
,
J.
2001
.
A 10-years fog climatology of Germany and the Alpine region based on satellite data—Preliminary results.
Proc. Second Int. Conf. on Fog and Fog Collection, St. John's, NF, Canada, Canadian International Development Agency, Canadian Meteor. and Oceanogr. Soc., and Environment Canada. 357–360
.
Bergot
,
T.
and
D.
Guedalia
.
1994
.
Numerical forecasting of radiation fog. Part I: Numerical model and sensitivity tests.
Mon. Wea. Rev.
122
:
1218
1230
.
Bott
,
A.
,
U.
Sievers
, and
W.
Zdunkowski
.
1990
.
A radiation fog model with a detailed treatment of the interaction between radiative transfer and fog microphysics.
J. Atmos. Sci.
47
:
2153
2166
.
Brown
,
R.
1980
.
A numerical study of radiation fog with an explicit formulation of the microphysics.
Quart. J. Roy. Meteor. Soc.
106
:
781
802
.
Bruijnzeel
,
L. A.
and
E. L.
Veneklaas
.
1998
.
Climatic conditions and tropical montane forest productivity: The fog has not lifted yet.
Ecology
79
:
3
9
.
Carson
,
R. B.
and
R. C.
Hardy
.
1963
.
On the forecasting of winter fog: A geographical approach.
J. Appl. Meteor.
2
:
351
357
.
Cavelier
,
J.
,
D.
Solis
, and
M. A.
Jaramillo
.
1996
.
Fog interception in montane forests across the central cordillera of Panama.
J. Trop. Ecol.
12
:
357
369
.
CIMIS,
2002
.
California irrigation management and information system—Data reports.
California Department of Water Resources. [Available online at http://www.cimis.water.ca.gov.]
.
Collet
,
J. L.
,
K. J.
Hoag
,
D. E.
Sherman
,
A.
Baton
, and
L. W.
Richards
.
1999
.
Spatial and temporal variations in San Joaquin Valley fog chemistry.
Atmos. Environ.
33
:
129
140
.
Court
,
A.
and
R.
Gerston
.
1966
.
Fog frequency in the United States.
Geogr. Rev.
56
:
543
550
.
Croft
,
P. J.
,
R. L.
Pfost
,
J. M.
Medlin
, and
G. A.
Johnson
.
1997
.
Fog forecasting for the southern region: A conceptual model approach.
Wea. Forecasting
12
:
545
556
.
d'Entremont
,
R. P.
1986
.
Low and midlevel cloud analysis using nighttime multispectral imagery.
J. Climate Appl. Meteor.
25
:
1853
1869
.
d'Entremont
,
R. P.
and
L. W.
Thompson
.
1987
.
Interpreting meteorological satellite images using a color-composite technique.
Bull. Amer. Meteor. Soc.
68
:
762
768
.
Duynkerke
,
P. G.
1991
.
Radiation fog: A comparison of model simulation with detailed observations.
Mon. Wea. Rev.
119
:
324
341
.
Ellrod
,
G.
1994
.
Detection and analysis of fog at night using GOES multispectral infrared imagery.
NOAA Tech. Rep. NESDIS-75, 22 pp
.
Ellrod
,
G.
1995
.
Advances in the detection and analysis of fog at night using GOES multispectral infrared imagery.
Wea. Forecasting
10
:
606
619
.
Ellrod
,
G.
2000
.
Proposed improvements to the GOES nighttime fog product to provide ceiling and visibility information.
Preprints, 10th Conf. on Satellite Meteorology and Oceanography, Long Beach, CA, Amer. Meteor. Soc., 454–456
.
Eyre
,
J. R.
,
J. L.
Brownscombe
, and
R. J.
Allen
.
1984
.
Detection of fog at night using Advanced High Resolution Radiometer (AVHRR) imagery.
Meteor. Mag.
113
:
266
271
.
Fisher
,
E. L.
and
P.
Caplan
.
1963
.
An experiment in numerical prediction of fog and stratus.
J. Atmos. Sci.
20
:
425
437
.
Fitzjarrald
,
D. R.
and
G. G.
Lala
.
1989
.
Hudson Valley fog environments.
J. Appl. Meteor.
28
:
1303
1328
.
Golding
,
B. W.
1993
.
A study of the influence of terrain on fog development.
Mon. Wea. Rev.
121
:
2529
2541
.
Guedalia
,
D.
and
T.
Bergot
.
1994
.
Numerical forecasting of radiation fog. Part II: A comparison of model simulation with several observed fog events.
Mon. Wea. Rev.
122
:
1231
1246
.
Gustafson
,
A. V.
and
S. E.
Wasserman
.
1976
.
Use of satellite information in observing and forecasting for dissipation and cloud formation.
Mon. Wea. Rev.
104
:
323
324
.
Hardwick
,
W. C.
1973
.
Monthly fog frequency in the continental United States.
Mon. Wea. Rev.
101
:
763
766
.
Holets
,
S.
and
R. N.
Swanson
.
1981
.
High-inversion fog episodes in central California.
J. Appl. Meteor.
20
:
890
899
.
Hunt
,
G. E.
1973
.
Radiative properties of terrestrial clouds at visible and infra-red thermal window wavelengths.
Quart. J. Roy. Meteor. Soc.
99
:
346
369
.
Kidder
,
S. Q.
and
T. H.
Vonder Haar
.
1995
.
Satellite Meteorology: An Introduction.
Academic Press, 446 pp
.
Kleespies
,
T. J.
1995
.
The retrievals of marine stratiform cloud properties from multiple observations in the 3.9-μm window under conditions of varying solar illumination.
J. Appl. Meteor.
34
:
1512
1524
.
Lala
,
G. G.
,
E.
Mandel
, and
J. E.
Jiusto
.
1975
.
A numerical evaluation of radiation fog variables.
J. Atmos. Sci.
32
:
720
728
.
Lee
,
T. F.
1987
.
Urban clear islands in California Central Valley fog.
Mon. Wea. Rev.
115
:
1794
1796
.
Lee
,
T. F.
,
F. J.
Turk
, and
K.
Richardson
.
1997
.
Stratus and fog products using GOES-8–9 3.9-μm data.
Wea. Forecasting
12
:
664
677
.
Mather
,
P. M.
1999
.
Computer Processing of Remotely-Sensed Images. 2d ed.
John Wiley and Sons, 292 pp
.
McLaren
,
S.
,
E.
McLaren
, and
M.
Mariotti
.
1988
.
Heterogeneous processes in the Po Valley radiation fog.
J. Geophys. Res.
93
:
11141
11151
.
Mead
,
J. B.
,
R. F.
McIntosh
,
D.
Vandemark
, and
C. T.
Swift
.
1989
.
Remote sensing of clouds and fog with a 1.4-mm radar.
J. Atmos. Oceanic Technol.
6
:
1090
1097
.
Menzel
,
W. P.
and
J. F. W.
Purdom
.
1994
.
Introducing GOES-1: The first of a new generation of geostationary operational environmental satellites.
Bull. Amer. Meteor. Soc.
75
:
757
782
.
Meyer
,
M. B.
and
G. G.
Lala
.
1990
.
Climatological aspects of radiation fog occurrence at Albany, New York.
J. Climate
3
:
577
586
.
Meyer
,
M. B.
,
G. G.
Lala
, and
J. E.
Jiusto
.
1986
.
FOG-82: A cooperative field study of radiation fog.
Bull. Amer. Meteor. Soc.
67
:
825
832
.
Meyer
,
W. D.
and
G. V.
Rao
.
1999
.
Radiation fog prediction using a simple numerical model.
Pure Appl. Geophys.
155
:
57
80
.
Moore
,
G. E.
,
C.
Daly
, and
M.
Liu
.
1987
.
Modeling of mountain–valley wind fields in the southern San Joaquin Valley, California.
J. Climate Appl. Meteor.
26
:
1230
1242
.
Musson-Genon
,
L.
1987
.
Numerical simulation of a fog event with a one-dimensional boundary layer model.
Mon. Wea. Rev.
115
:
592
607
.
NCDC,
cited 2002
.
National virtual data system.
National Climatic Data Center, NOAA, Department of Commerce, Asheville, NC. [Available online at http://nndc.noaa.gov/?home.shtml.]
.
Oke
,
T. R.
1988
.
Boundary Layer Climates.
Routledge, 435 pp
.
Peace
,
R. L.
1969
.
Heavy-fog regions in the conterminous United States.
Mon. Wea. Rev.
97
:
116
123
.
Pilie
,
R. J.
,
E. J.
Mack
,
W. C.
Kocmond
,
W. J.
Eadie
, and
C. W.
Rogers
.
1975a
.
The life cycle of valley fog. Part II: Fog microphysics.
J. Appl. Meteor.
14
:
364
374
.
Pilie
,
R. J.
,
E. J.
Mack
,
W. C.
Kocmond
,
C. W.
Rogers
, and
W. J.
Eadie
.
1975b
.
The life cycle of valley fog. Part I: Micrometeorological characteristics.
J. Appl. Meteor.
14
:
347
363
.
Pinnick
,
R. G.
,
D. L.
Hoihjelle
,
G.
Fernandez
,
E. B.
Stenmark
, and
J. D.
Lindberg
.
1978
.
Vertical structure in atmospheric fog and haze and its effects on visible and infrared extinctions.
J. Atmos. Sci.
35
:
2020
2032
.
Putsay
,
M.
,
J.
Kerenyi
,
I.
Sebok
,
P.
Nemeth
, and
M.
Dioszeghy
.
2001
.
Nighttime fog and low cloud detection in NOAA-16 AVHRR images and validation with ground observed SYNOP data and radar measurements.
Preprints, The 2001 EUMETSAT Meteorological Satellite Data Users' Conf., Ankura, Turkey, Euopean Organisation for the Exploitation of Meteorological Satellites, 365–373
.
Rao
,
P. K.
,
S. J.
Holmes
,
R. K.
Anderson
,
J. S.
Winston
, and
P. E.
Lehr
.
Eds.,
.
1990
.
Weather Satellites: Systems, Data, and Environmental Applications.
Amer. Meteor. Soc., 503 pp
.
Rawlings
,
F.
and
J. S.
Foot
.
1990
.
Remotely sensed measurements of stratocumulus properties during FIRE using the C130 aircraft multi-channel radiometer.
J. Atmos. Sci.
47
:
2488
2504
.
St. Jean
,
D.
1997
.
Using varied enhancement curves to better evaluate fog and status with RAMSDIS.
U.S. Department of Commerce, Western Region Tech. Attachment 97-40, 9 pp
.
Suckling
,
P. W.
and
M. D.
Mitchell
.
1988
.
Fog climatology of the Sacramento urban area.
Prof. Geogr.
40
:
186
194
.
Suomi
,
V. E.
,
R.
Fox
,
S. S.
Limaye
, and
W. L.
Smith
.
1983
.
McIDAS III: A modern interactive data access and analysis system.
J. Climate Appl. Meteor.
22
:
766
778
.
Swenson
,
S. E.
2002
.
Crashes clog highways.
Bakersfield Californian, 3 January
.
Welch
,
R. M.
and
S. K.
Cox
.
1986
.
Prediction of quasi-periodic oscillations in radiation fogs. Part I: Comparison of simple similarity approaches.
J. Atmos. Sci.
43
:
633
651
.
Wetzel
,
M. A.
,
R. D.
Borys
, and
L. E.
Xu
.
1996
.
Satellite microphysical retrievals for land-based fog with validation by balloon profiling.
J. Appl. Meteor.
35
:
810
829
.
Yarnal
,
B.
1994
.
Synoptic Climatology in Environmental Analysis: A Primer.
John Wiley and Sons, 256 pp
.

Footnotes

Corresponding author address: Dr. S. Jeffrey Underwood, Department of Geography, 4442 Faner Hall, Southern Illinois University, Carbondale, IL 62901-4514. jeffreyu@siu.edu