Desert dust blown from the arid regions of the Colorado Plateau and Great Basin contributes to earlier snowmelt in the Upper Colorado River Basin. A recent study indicates that this premature snowmelt reduces the annual flow of the Colorado River by five percent.
Based on hydrological modeling, the study co-authored by Brad Udall and Jeffrey Deems, based at the Western Water Assessment, and Tom Painter, snow hydrologist at NASA’s Jet Propulsion Laboratory, indicates that between 1915 and 2003 snowmelt has occurred on average three weeks earlier on mountains in the Upper Colorado River Basin than it did prior to the arrival of Anglo-Saxon settlers in the 1850s.
Dust accumulated on the snowpack decreases snow’s albedo – its capacity to reflect solar radiation. Because of dust’s dark color it absorbs more sunlight, leading to accelerated melting. This premature snowmelt extends the growing season of plants that under historical conditions would be covered by snow until later in the season. Evapotranspiration, the process whereby water moves from plants and soil into the atmosphere, results in the loss of some 750, 000 acre-foot of water each year. This is sufficient to supply Los Angeles for 18 months.
This yearly deficit has been unwittingly incorporated into the allocation of the Colorado River, which serves 27 million people in seven states and Mexico. The study indicates the river is over allocated by ten percent.
Jeffrey Deems says that even if mitigation measures, such as improved land management practices, curtail the annual water loss it won’t be a “silver bullet.”
The predicted future effects of climate change in the Southwest need to be carefully considered. Also, Deems notes that in 2009 and 2010 there was a particularly high dust load on the snow.
I spoke with Deems to get some insights into the study and the effect of dust on snow. (I’m running this as a QnA).
BB: I’ve read that average temperatures have increased in the Southwest during the twentieth century and the winter storm track is moving northward. Is there any way this warming is contributing to snow melting earlier?
JD: It definitely is doing that. The impacts of climate change and dust can both be characterized as forcing: there’s something extra in the system beyond its average state that’s pushing the response. The dust forcing is currently vastly stronger than the climate forcing.
Climate change impacts have a direct effect on snowmelt and snow accumulation. They also have a direct effect on soil moisture and vegetation in the desert. It’s not so much the direct climate forcing on the snow that’s the issue, but the climate forcing in the desert that has a bigger impact on snowmelt.
BB: This dust comes from the Colorado Plateau and northwest New Mexico, northeast Arizona and southern Utah. Is climate change producing more dust in these regions?
JD: We’re working on that. The easy answer is yes, but I think that’s too easy of an answer. We’re in an extended period of drought in the Southwest. Drought definitely impacts vegetation growth, soil moisture availability, re-growth of crusts on the soil surface, and also invasive plant species.
There are definitely impacts happening out there that are no doubt influenced by climate warming. Now, that will very quickly devolve into a detailed argument. How would the duration and magnitude of the current drought be changed if we were still at pre-industrial levels? That’s a very difficult question to answer.
All we can really say is that there are impacts of climate change being felt in the dust emission regions that are likely contributing to enhanced emissions, but we don’t know that. What is worrisome is the potential for further impacts of climate change on the vegetation and land condition in the desert, which could dramatically enhance dust emissions.
In 2009 and 2010 we’ve seen quite a lot of dust on the snowpack. The current conditions scenario in our modeling study closely matched conditions we observed in 2003 to 2008. They did not represent the severe dust loading of the past two years. Our current work is to replicate the same study with an enhanced dust scenario that matches 2009 and 2010.
If you look at 2008 and 2009, we had the same amount of dust events after March 1, but 2009 had over four times the dust concentration in the snowpack. In 2008 snow melted out about 25 days earlier than it would have in a pristine snowpack. In 2009 we were on the order of 50 days.
There’s something about the land conditions out in the source regions that allowed the same number of storms to deposit a lot more dust. Maybe the storms were more intense. There’s a whole chain of processes involved, from dust being available for transport, to getting picked up, carried and deposited in different locations, and then affecting snowmelt.
BB: The study indicates that in 1920 dust loading was six times that of pre-1850 conditions. Since then it has averaged a fivefold increase on these condition. Is it possible that there could be a future event where it exceeds the sixfold level of dust?
JD: Yes. You know, I haven’t actually done any math on it, but it’s likely the last two years have greatly exceeded that. That sixfold increase comes from Jason Neff’s work. He and his team took lake sediment cores in some alpine lakes in the San Juans. The sediment accumulation rate in the lakes is a good barometer of dust deposition. Not to a time resolution of an individual year, though, but it’s relatively high, kind of sub-decadal.
Round about the mid 1800s we see a dramatic jump – 600 percent jump in dust accumulation. That drops back again in the 1930s coincident with, and this is just circumstantial, the Taylor Grazing Act, which reduced the number of grazing animals on public lands in this region.
Despite dust emission being a natural process in these regions, that disturbance of the desert soil is really critical to having large amounts of dust. By disturbance I mean breaking up of the desert crust. When well developed it makes the surface virtually impervious to wind erosion.
BB: Is the crust being disturbed by agriculture and grazing?
JD: Grazing, agriculture, energy exploration, road building, and recreational users. There have been some efforts to remove certain types of vegetation and plant others to make the forage more palatable for grazing animals. Basically, anything that busts up the crust and does it over a large enough area is going to have an effect.
The lake sediments suggest that if you remove the disturbance the crust has the potential to regenerate. Jason Neff’s work at sites at the Dugout Ranch in Canyonlands National Park, looking at dust emissions of currently grazed land, land that was grazed thirty years ago or more but hasn’t been grazed since, and land that hasn’t been grazed at all, leads us to suggest that improved land management could reduce the dust impacts on snowmelt. You could characterize it as getting water back. We’ll see; it may just be a partial offset for climate change impacts or some insurance against climate change impacts directly on the snow and in the source regions.
BB: Climate models predict a seven to twenty percent reduction in Colorado River flow by 2050 due to human-induced climate change. If you compound this with an average five percent loss of water each year, what is the net effect?
JD: It’s tempting to add those together, but there’s a nuance to it. The biggest impact on the timing of snowmelt due to dust comes in bigger snow years or in areas that receive more snow. The amount of snow on the ground determines in large part the magnitude of the impact on timing, which controls the evaporative processes. If climate impacts reduce the snow accumulation, the fractional dust impact will be reduced. That’s not necessarily good news: the dust will still have an impact but the actual number may be reduced. It would be foolish not to take advantage of any potential to mitigate the dust mechanisms.
This isn’t a remove all humans from the equation kind of solution, but identifying hotspots for emission and changing land management practices…
BB: And taking the study results into consideration and working them into land management practices?
JD: It’s easy to wrap everything up in the term “land management practices,” but a lot of the dust emission is coming from the Navaho nation lands. The Navaho people have been socially and economically marginalized, to put it kindly. Range conditions show it. It’s not as simple as saying you need to change your land management practices. It’s going to require social and economic engagement.
BB: The Colorado River Compact and Law of the River legislation is actually based on a misunderstanding of the river’s normal conditions. Why is that?
JD: The time period over which they averaged the flow in order to divide up the river was an anomalously high flow period. The river is over allocated. Currently, the upper basin states are not using their full allocation and that “extra water” has been fairly reliably delivered to the lower basin and kept Lake Powell and Mead somewhat full.
BB: Say mitigation measures do cut out this five percent yearly loss. Would it be enough to fill the lakes back up?
JD: All the main stem storage on the Colorado River adds up to the capacity of four years of river flow. If we change things by five percent of annual flow it’s not going to fill things up right away, but that five percent is twice Las Vegas’ current water allocation. It’s about half the annual shortfall in Lake Mead.
If we were somehow able to “regain” the five percent it’s not an inconsequential amount of water, but it’s also not a silver bullet. Now we’re good, don’t worry about climate change – that’s not the thing at all. Anything we can do to restore range conditions in the dust emission areas and reduce dust impacts on snowmelt will be helpful.