@KarlTheFog Has Been Mapped!

Alicia torregrosa, physical Scientist

Within the world of mapping, clouds are a pesky interference to be removed from satellite remote sensed imagery.  However, to many of us, that is a waste of pixels. Cloud maps are becoming increasingly valuable in the quest to understand land cover change and surface processes. In coastal California, the dynamic summertime interactions between air masses, the ocean, and topography result in blankets of fog and low clouds flowing into low lying areas of the San Francisco Bay Area. The low clouds and fog advected from the Pacific bring moisture and shade to coastal ecosystems. This acts to reduce temperatures and evapotranspiration stress during the otherwise arid Mediterranean climate season, in turn impacting vegetation distribution, irrigation needs, and urban energy consumption.

Figure 1. Landsat, May 22, 1991

The Landsat image of May 22, 1991 shows the marine stratus and stratocumulus cloud layer that moved into the San Francisco Bay-Delta and Monterey Bay. Several cloud patterns can be seen in this image: the eddy-like spiral to the west of the Golden Gate, the darker linear cloud feature that parallels the coast down to Monterey Bay, and fog funneling from Monterey Bay inland. Researchers are using synoptic meteorology, aerosol chemistry, cloud thermodynamics, microphysics, and physical oceanography to better understand the processes behind these patterns. This research will improve the representation of clouds in global circulation and climate models.

To early natural history scholars, the distribution of iconic coastal California vegetation types such as redwood forests, maritime chaparral, and endemic species such as Torrey and Monterey pines was explained by their spatial coincidence with the summertime fog belt zone. Although the term “fog belt” has long been commonly used, it was not quantified with the precision or accuracy needed by current ecosystem models. The absence of a fog belt map for scientific purposes was particularly puzzling to local scientists and would be to anyone watching the local evening news. We’ve gotten used to seeing fog maps as part of the weather segment. Doesn’t this mean fog belt maps exist?

Unfortunately, fog maps are devilishly difficult to generate for two reasons. First, coastal fog and low cloud cover are continuously moving across the landscape, sometimes covering more ground at night than during the day. Images from Landsat that give us beautiful daytime snapshots of coastal fog along a 90 km-wide path every 16 days, as shown in figure 1, are insufficient to quantify the full spatiotemporal extent of fog and low cloud cover (FLCC). Temporal coverage is much higher from the National Weather Service geostationary operational environmental satellite (GOES) that relays images every 15 minutes. Figure 2 shows an FLCC map derived from 26,000 hourly GOES images from June, July, August, and September (1999 – 2009). 

Figure 2. San Francisco Bay Area Fog Belt Zones

The Fog and Low Cloud Cover (FLCC) Map of the San Francisco Bay Area depicts the average hours of FLCC received during June, July, August, and September 1999 – 2009 in contour intervals of 30 minutes.   The averages were generated as a raster map from hourly night and day GOES images. The raster was interpolated into contours for visualization purposes. Note:  a high resolution 18" x 36" jpg version of this map is available in this issue of the BayGeo Journal map gallery.

The second challenge to generating a fog map lies in the vertical dimension. One step in processing the GOES imagery to generate the FLCC map was differentiating and removing high clouds from low clouds. The temperature difference between colder high cirrus and warmer low stratus or stratocumulus is calculated using the shortwave and thermal infrared bands from the GOES Imager sensor. Pixels are processed against background surface thermal thresholds to be classified into no cloud, low cloud, or high cloud categories. However, as the singer Joni Mitchell tells us, we can look at clouds from up and down and really not know clouds at all. The satellite view is exclusively down-looking, which means we cannot tell if the low cloud is actually touching the ground. If it is touching, it is fog, if not, it is just a low cloud. 

To beachgoers who had hoped for a sunny day, telling the difference between summertime low clouds and fog is hardly worth the effort, but to the redwood trees and streams that benefit from fog drip, the difference means as much as a 200% increase in stream flow. Some years, such as 2011, are very foggy, whereas others, such as 2014, are not. The interannual, intraseasonal (monthly), and diurnal variability of FLCC is ecologically significant and therefore of interest to natural resource managers. During times of extremely low streamflow Salmon Protection and Watershed Network (SPAWN) volunteers conduct fish rescues relocating fish from pools going dry to reaches with perennial flow. For long-term ecological restoration purposes, FLCC statistics that identify areas with the highest stability can be insightful. Understanding which riparian areas have the highest FLCC under most interannual conditions could affect decisions about what species to plant in a given riparian corridor.

Figure 3. Fog and low cloud clover (FLCC) by average month-hour

The hourly sequence of average percent cover of fog and low cloud (FLC) at 4 different hours (8, 9, 10, and 11 AM PDT) for July and September in 2001.  On average, by 10 AM most of the Salinas Valley (a3) is clear of FLC.  Many other average hourly patterns can be found in these sequences such as increasing FLCC offshore by 11 AM and increasing FLCC in the easternmost edge of the map (Central Valley) in July.

Meteorological questions also abound. What time of day, on average, does FLCC disappear? In some areas nighttime fog events are most common, in other areas the fog is present through most of the morning. Aggregating the 26,000 hourly GOES images by month-hours (the average monthly FLCC for each hour of a 24-hour period), as shown in figure 3, provides a high resolution temporal slice to compare changes on an hourly basis without getting too overwhelmed by big data.   

Fog holds many intriguing mysteries beyond the question of ‘Why are some years or hours foggier than others?’ Scientists hypothesize the most probable and strongest drivers are ocean cycles and atmospheric inversion strength. To improve understanding about coastal fog, scientists from governmental, academic, and non-profit organizations from coastal foggy areas of the world are banding together to co-locate ground-based fog monitoring and research instruments. Analyzing collocated data will help to explore the difference in fog events in terms of fog water content, aerosol composition (salts, microbes, plankton, nutrients, pollution, and toxins are all known to be cloud condensation nuclei), contribution to groundwater, solar radiation dynamics, and much more. To join in this research or read more check out the Pacific Coastal Fog Project at http://geography.wr.usgs.gov/fog/. To access the data described in this article check out the California Climate Commons webpages on coastal fog at http://climate.calcommons.org/projectlink/fog.

As Karl the Fog (twitter handle @KarlTheFog) watchers will tell you, next time you do/don't want to be in fog, just wait a few hours and it will change, but if you are impatient just move a few miles and you'll find/lose some sun. 

Alicia Torregrosa can be reached at atorregrosa@usgs.gov.


Spring 2016 Volume 9 Issue 1