Unseen Dust Emission and Global Dust Abundance: Documenting Dust Emission from the Mojave Desert (USA) by Daily Remote Camera Imagery and Wind‐Erosion Measurements

A large gap in understanding the effects of atmospheric dust at all spatial scales is uncertainty about how much and whence dust is emitted annually. Direct digital recording of dust emission at high spatial and temporal resolution would together with periodic flux measurements support improved estimates of local-scale dust flux. Remote camera recording of dust-emitting settings on and around Soda Lake (Mojave Desert) was conducted every 15 min during daylight between 10 November 2010 and 31 December 2016. Examination of 135,000 images revealed frequent dust events characterized as dust days, 68 per year on average, when dust plumes obscured mountains beyond areas of dust generation. Dust was not observed from examination of satellite retrievals (Moderate Resolution Imaging Spectroradiometer and Geostationary Operational Environmental Satellite System) during six cloudless days of extreme emission. Estimated masses of emitted dust fractions were made from measured horizontal mass flux, sediment particle sizes, and areas of dust generation. Between April 2000 and June 2013, nearly 4 Tg/year of dust (particles < 63 mu m) were emitted across the study area. Higher rates (about 7 Tg/year) were estimated following the December 2010 Mojave River flood that deposited sediment across the lake basin. Within the Mojave and Great Basin deserts of western North America, many settings akin to those at Soda Lake similarly emit dust that is rarely detected in satellite retrievals, implying that local and regional dust emissions from western North America are far underestimated and that, by extension to relatively small dust-source areas across all drylands, global dust emissions may also be greatly underestimated. Plain Language Summary Small mineral particles in the air (atmospheric dust) affect many aspects of human and other environments-air quality and public health, delivery of nutrients across wide swaths of ocean and land, melting of snow and ice, and air temperature where and when dust plumes move across the globe. We cannot fully understand these effects until we know better how much dust is emitted from the land surface each year and the locations of dust-source areas where strong winds loft it into the atmosphere. Mostly from measurements and images from satellites, current ways to detect dust and estimate the annual mass of global dust can sense large emissions from dust storms (most global emission estimates are 1000-4000 million metric tons per year), but they likely miss the full ranges of places and amounts of such emission. To examine how much dust might go unseen and uncounted by satellite methods, we installed a remote camera to visually record dust emission from a small dust-source area, Soda Lake in the Mojave Desert, California. Images were taken every 15 min during daylight between 10 November 2010 and 31 December 2016, yielding 135,000 images. Dust emission was seen in the images during nearly 20% of days. Dust was never observed, however, from examination of satellite images at simultaneous periods of large dust storms seen from the ground. From periodic weighing of dust collected on the ground between April 2000 and June 2013 and an estimate of the area of dust generation, we roughly calculated the masses of emitted dust to be between 4 and 7 million metric tons of dust each year. Within the Mojave and Great Basin deserts of western North America, hundreds of settings akin to those around Soda Lake emit dust that is rarely detected by satellite, strongly implying that local and regional dust emissions from western North America are far underestimated and that, by extension to relatively small dust-source areas across all drylands, global dust emissions are probably greatly underestimated.