AI News, Machine learning predicts new details of geothermal heat flux beneath the Greenland Ice Sheet

Machine learning predicts new details of geothermal heat flux beneath the Greenland Ice Sheet

Among the key findings: Greenland has an anomalously high heat flux in a relatively large northern region spreading from the interior to the east and west.

'Heat that comes up from the interior of the Earth contributes to the amount of melt on the bottom of the ice sheet -- so it's extremely important to understand the pattern of that heat and how it's distributed at the bottom of the ice sheet,' said Soroush Rezvanbehbahani, a doctoral student in geology at the University of Kansas who spearheaded the research.

Rezvanbehbahani and his colleagues have adopted machine learning -- a type of artificial intelligence using statistical techniques and computer algorithms -- to predict heat flux values that would be daunting to obtain in the same detail via conventional ice cores.

Using all available geologic, tectonic and geothermal heat flux data for Greenland -- along with geothermal heat flux data from around the globe -- the team deployed a machine learning approach that predicts geothermal heat flux values under the ice sheet throughout Greenland based on 22 geologic variables such as bedrock topography, crustal thickness, magnetic anomalies, rock types and proximity to features like trenches, ridges, young rifts, volcanoes and hot spots.

'We have a lot of data points from around the Earth -- we know in certain parts of the world the crust is a certain thickness, composed of a specific kind of rock and located a known distance from a volcano -- and we take those relationships and apply them to what we know about Greenland,' said co-author Leigh Stearns, associate professor of geology at KU.

The authors found the five most important geologic features in predicting geothermal flux values are topography, distance to young rifts, distance to trench, depth of lithosphere-asthenosphere boundary (layers of the Earth's mantle) and depth to Mohorovičić

Machine Learning Model Forecasts Geothermal Heat Flux Under Greenland’s Ice

A new prediction model using machine learning forecasts a slightly elevated heat flux upstream of several fast-flowing glaciers in Greenland, including Jakobshavn Isbrae, the fastest moving glacier on Earth.

Instead of expensive field surveys, we try to do this through statistical methods.” The team found that Greenland has an anomalously high heat flux in a relatively large northern region spreading from the interior to the east and west, while southern Greenland has a relatively low geothermal heat flux, corresponding with the extent of the North Atlantic Craton—a stable portion of one of the oldest extant continental crusts on Earth.

The machine learning system uses all available geologic, tectonic and geothermal heat flux data for Greenland, as well as geothermal heat flux data from around the globe to predict geothermal heat flux values under the ice sheet throughout Greenland.

“We try to incorporate as many geologic data sets as we can rather than assuming one is the most important.” The researchers found the five most important geologic features in predicting geothermal flux values are topography, distance to young rifts, distance to trench, depth of lithosphere-asthenosphere boundary (layers of the Earth's mantle) and depth to Mohoroviči discontinuity (the boundary between the crust and the mantle in the Earth).

High geothermal heat flux in close proximity to the Northeast Greenland Ice Stream

A recent notable increase in surface speed is observed in the northeastern region of Greenland originating from the center of GIS and exiting through the ice-stream outlet glaciers;

During summer, hydrographic conditions in the fjord are characterized by highly stratified surface waters and a relatively wide range of salinities (17 to 33.4) and temperatures (−1.6 °C to 8 °C)19, while winter conditions are weakly stratified with near ocean freezing temperatures (c.

On a sea ice expedition to the fjord in March 2005 we noticed that water temperatures started to increase in the deepest part of Tyrolerfjord from ~250 m depth toward the bottom in the deep basin and this temperature gradient has also been present in the subsequent measurements from this area since (Fig. 2).

The heat source for this deep heating was analyzed further by considering all profiles of potential temperature (θ) and salinity (S) from the period 2005–2015 in the deepest part of the fjord, i.e.

Subsequently there were no indications of further major bottom water intrusions or deep convection in the period from 2006–2015 where θ and S below 240 m in average increased by 0.016 °C yr−1 and decreased by 0.03 psu per year, respectively.

During this period the highest potential temperature below 250 m was observed close to the bottom in the “Dybet” section and this temperature increase was also manifested as a concave shape of the annual measurements in the θS-curve (Fig. 2c).

The annual temperature increase of bottom water below 240 m depth could either be explained by a gradual bottom water renewal in the fjord system or due to internal processes in the “Dybet” section, i.e.

Thus, the potential for gradual bottom water renewal was analyzed by comparing all profiles in the “Dybet” section from each year with corresponding profiles of coastal water masses outside the fjord (data not shown).

The sill depth of 45 m is a barrier for bottom water exchange and to account for a potential isopycnal heave of ~15 m, we compared coastal water mass characteristics down to 60 m with the densities below 240 m depth in the “Dybet” section.

the year with a major bottom water renewal, all profiles showed that the density of water in the “Dybet” section below 240 m depth was higher than coastal water masses down to 60 m depth.

However, measurements from the entire period show that the highest density and salinity occurred in late summer, and this support the conclusion above based on water mass analysis of measurements from July-September.

During this period an enhanced surface transport was recorded out of the fjord and a layer between 40–140 m depth was flowing into the fjord and supplying the interior with cool and saline water from the polynya.

An annual potential temperature anomaly (Δθ = θ − <θ>) was defined by considering all profiles of θ in the Dybet section below 240 m in relation to the annually averaged θ of all profiles between 260–280 m (<θ>), i.e.

The spatial distribution of Δθ for each year showed that maximum temperature anomalies were always found close to the bottom and the averaged distribution for the period 2006–2015 also showed a significant temperature increase of up to 0.02 °C (Fig. 3a).

Temporal change of a tracer (φ, e.g S, θ or the heat content H) below the interface depth (Di = 240 m) was assumed to be due to vertical turbulent mixing in the period 2006–2015 and, thereby, described by the one-dimensional conservation equation: where t and z is time and the vertical coordinate, respectively, and k

Assuming that the total temperature change below 240 m depth originates from the bedrock heat flux below the deep basin, this represents a total heat flux of 110 ± 36 mW m−2 in the period 2006–2015 (Fig. 4c).

A more likely out-flowing temperature range of ~0 °C, corresponding to glacial meltwater, would imply a volume flux of more than 1 m3 s−1 and lead to a salinity change of ~50 times larger than observed.

Therefore, the GHF can either be explained as a conductive heat flux or as due to a weakly outflowing and warm deep (>16 °C) meteoric water source and, in this case, the calculated GHF would be a lower bound because less of the temperature increase in Dybet would be due to vertical turbulent transports in the water column.

Elevated GHF providing heat for basal melt and, thereby, increasing glacier sliding over the bedrock could be a contributing factor although local ice geometrical settings, subglacial hydrology and the mechanical properties of the ice-bedrock interface are also important modulating factors.

Journal of Geophysical Research: Earth Surface

[45] In this section, we examine the effects of the different GHF forcings integrated over

the GIS history on the modeled present‐day topography and ice temperature distribution,

and compare the modeled and observed ice thicknesses and temperatures at the locations

[46] In Figure 3 the differences between the modeled and observed present‐day ice thicknesses [Bamber et al., 2001;

discrepancies between the modeled and observed thicknesses, exceeding 1km at the

margins of the GIS and in two large areas in the most northern and eastern parts of

temporal climate variations and ice sheet margin dynamics, which are responsible for

ice margin areas and in those areas with large horizontal gradients of bedrock and

regions is associated with higher uncertainties due to the strong influence of small‐

and large‐scale variations in bedrock topography on the GHF values [van der Veen et al., 2007].

[47] We can observe that all four simulations equally fail to reproduce the observed ice

between the observed and modeled thicknesses vary between 100 and 300m, whereas the

misfit reaches 400 to 600m on the eastern side, including the location of the Dye

the four runs, the MDS provides the best fit to the divide thicknesses of the present‐day

[49] The present‐day basal temperatures BPMP (below pressure‐melting point, see section 2.1), T′, computed by the four simulations are shown in Figure 4.

site still lying at the margin of the cold‐ice area surrounding the southern dome

should have a cold base in compliance with the extremely low basal temperature values

as already suggested by the analysis of the basal temperature evolution (see section 3.1).

and basal temperatures, T, at all ice core locations are too high compared to the observed values, with the

[51] By contrast, the STS produces the lowest basal temperatures BPMP, T′, and the smallest temperate ice area among the four runs.

the marginal areas, basal temperatures BPMP computed by the STS lie within the range

[52] The basal temperatures, T, from the MDS are about 5°C too high in the area of the GRIP and GISP2 stations (Table 2), but are close to the observed values at the NGRIP and CC and in agreement with

base area (Figure 4c) occupying more than 56% of the present‐day GIS‐covered territory, whereas large

[53] The U63S produces the best agreement of the four runs with the observed basal temperatures

[54] In agreement with the existing data and estimations [Fahnestock et al., 2001;

[55] In Figure 5, we present modeled versus measured temperatures, T, at the locations of the GISP2, GRIP, NGRIP and Dye 3 ice cores (see Figure 1 for their locations) [Clow et al., 1996;

Johnsen et al., 1995] and differences between modeled and measured temperature profiles at the locations

temperatures between the ice base and 2000m elevation above the bedrock, with the

profiles, being 1.6°C too‐low at the surface due to the lapse rate (0.79°C per 100m)

All modeled results show significant differences with respect to the observed vertical

[56] At the NGRIP site (Figure 5c), the STS produces ice almost 200m thicker, with the surface and basal temperatures

[57] Finally, at the Dye 3 location (Figure 5d), all modeled temperature profiles closely coincide with each other, although the

Earth's internal heat drives rapid ice flow and subglacial melting in Greenland

Earth's internal heat drives rapid ice flow and subglacial melting in Greenland Greenland's lithosphere has hot depths which ..

Heat loss from Earth's interior triggers ice sheet slide towards the sea

Heat loss from Earth's interior triggers ice sheet slide towards the sea Greenland's ice sheet is becoming smaller and smaller. The melting takes place with ...

Study bolsters theory of heat source under Antarctica

Study bolsters theory of heat source under Antarctica A new NASA study adds evidence that a geothermal heat source called a ..

Lecture - 35 Geothermal Energy

Lecture Series on Energy Resources and Technology by Prof.S.Banerjee,Department of Electrical Engineering, IIT Kharagpur. For more details on NPTEL visit ...

NASA | Glacial Ice Loss: Greenland and Antarctica

In the first comprehensive satellite study of its kind, a University of Colorado Boulder-led team used NASA data to calculate how much Earth's melting land ice is ...

Mantle Plumes, Not Global Warming, Cause of Ice Melt

by Mitch Battros April 4th 2016 To understand Greenland's ice of today - researchers have to go far back into Earth's history. The island's lithosphere has hot...

Heat loss from the Earth triggers ice sheet slide towards the sea

Today, in the esteemed journal Scientific Reports, researchers from the Arctic Research Centre, Aarhus University, and the Greenland Institute of Natural ...

Scientists have found evidence of another natural factor Earth's internal heat melts ice sheets

Scientists have found evidence of another natural factor Earth's internal heat melts ice sheets At a time when polar ice caps and glaciers around the world are ...

Friendly Guide to Climate Change - and what you can do to help #everytoncounts

This video is a short, sweet, and pragmatic summary of climate change - what the problem is, why, and what you can actually do about it. VOICE ...

NASA | Lakes On A Glacier

How deep is that icy blue water on Greenland's ice sheet? Dr. Allen Pope, of the National Snow and Ice Data Center, is using data from the NASA/USGS ...