Search

The thermochemical nature of the lower mantle, and the two Large Low Shear-wave Velocity Provinces (LLSVPs) in particular, remains a topic of active debate. Exclusively imaging seismic velocity anomalies only provides limited ability to distinguish between thermal and compositional origins. Here we use whole Earth oscillations, or normal modes, to study 3D variations in mantle shear wave- (vs), compressional wave- (vp) and bulk sound velocity (vc), density (rho) and shear attenuation (qmu). These observations provide new constraints on the presence of lower mantle chemical heterogeneity. Here, we compare the two-step splitting function inversion method to the less frequently used, computationally more expensive one-step direct spectrum inversion. In theory, the one-step inversion suffers less from non-uniqueness and only requires regularization once. In practice, we find that the average spectral misfits for the one-step inversion are always lower. The ratio between vs and vp anomalies obtained from their joint inversion, proposed to be an indicator of chemical heterogeneity when exceeding a threshold predicted by mineral physics, varies significantly between the two inversion methods. The method of computing the ratio is just as important. We obtain ratios exceeding the threshold in the lower mantle only when dividing the root mean square amplitudes of our vs and vp models, although lower than some previous studies suggest. However, by taking the median ratio from a grid at each depth, we barely exceed the threshold. Instead of relying on these 1D representatives of the ratio, we infer chemical heterogeneity in certain depth ranges based on a wide spread in distributions of vs, vp anomalies and their ratio. Another constraint on the presence of chemical heterogeneity comes from the anti-correlation of vs and vc structure, found in many previous studies. We find (de-)correlation in the lower mantle in joint vs and vc inversions, with only slightly negative correlation for the two-step inversion. We shift towards more negative correlation values when extracting vc from our previously obtained vs and vp models. Lower mantle density structure has remained elusive and controversial in recent decades. In joint inversions for vs, vp, rho and discontinuity topography, we show that a basal layer of excess density is located underneath part of the otherwise lighter-than-average LLSVPs, reconciling previous observations. This dense layer is more robust in the one-step inversion and does not result from high-amplitude ghost patterns that may have plagued earlier normal mode studies. Compositional variations such as iron enrichment may explain this dense but seismically slow layer. Finally, we obtain preliminary results for imaging lower mantle shear attenuation in joint vs and qmu one-step inversions. We first select the best way of inverting for 3D qmu in synthetic tests, before applying this method to real data. Mode selection appears to be crucial, and we also need to properly account for detailed elastic structure to minimize leakage into anelastic structure. The preliminary 3D qmu model shows low attenuation in LLSVPs and high attenuation in the surrounding lower mantle, possibly indicating a dominant role for grain size.

The thermochemical nature of the lower mantle, and the two Large Low Shear-wave Velocity Provinces (LLSVPs) in particular, remains a topic of active debate. Exclusively imaging seismic velocity anomalies only provides limited ability to distinguish between thermal and compositional origins. Here we use whole Earth oscillations, or normal modes, to study 3D variations in mantle shear wave- (vs), compressional wave- (vp) and bulk sound velocity (vc), density (rho) and shear attenuation (qmu). These observations provide new constraints on the presence of lower mantle chemical heterogeneity. Here, we compare the two-step splitting function inversion method to the less frequently used, computationally more expensive one-step direct spectrum inversion. In theory, the one-step inversion suffers less from non-uniqueness and only requires regularization once. In practice, we find that the average spectral misfits for the one-step inversion are always lower. The ratio between vs and vp anomalies obtained from their joint inversion, proposed to be an indicator of chemical heterogeneity when exceeding a threshold predicted by mineral physics, varies significantly between the two inversion methods. The method of computing the ratio is just as important. We obtain ratios exceeding the threshold in the lower mantle only when dividing the root mean square amplitudes of our vs and vp models, although lower than some previous studies suggest. However, by taking the median ratio from a grid at each depth, we barely exceed the threshold. Instead of relying on these 1D representatives of the ratio, we infer chemical heterogeneity in certain depth ranges based on a wide spread in distributions of vs, vp anomalies and their ratio. Another constraint on the presence of chemical heterogeneity comes from the anti-correlation of vs and vc structure, found in many previous studies. We find (de-)correlation in the lower mantle in joint vs and vc inversions, with only slightly negative correlation for the two-step inversion. We shift towards more negative correlation values when extracting vc from our previously obtained vs and vp models. Lower mantle density structure has remained elusive and controversial in recent decades. In joint inversions for vs, vp, rho and discontinuity topography, we show that a basal layer of excess density is located underneath part of the otherwise lighter-than-average LLSVPs, reconciling previous observations. This dense layer is more robust in the one-step inversion and does not result from high-amplitude ghost patterns that may have plagued earlier normal mode studies. Compositional variations such as iron enrichment may explain this dense but seismically slow layer. Finally, we obtain preliminary results for imaging lower mantle shear attenuation in joint vs and qmu one-step inversions. We first select the best way of inverting for 3D qmu in synthetic tests, before applying this method to real data. Mode selection appears to be crucial, and we also need to properly account for detailed elastic structure to minimize leakage into anelastic structure. The preliminary 3D qmu model shows low attenuation in LLSVPs and high attenuation in the surrounding lower mantle, possibly indicating a dominant role for grain size.

Liermåsan like nord for Bjørkelangen er et torvuttak som er under avslutning på ca 1,2 km2. Lierelva renner forbi torvuttaket og rett inn i Bjørkelangen sentrum. Området har problemer med flom, og rapporten har utredet potensialet og metoder for hvordan Liermåsan kan brukes til fordrøyning for å dempe flommene. Areal- og volumberegninger viser at Liermåsan kan lagre fra ca 165000 m3 til ca 2400000 m3 etter hvor omfattende tiltak som gjøres. For å redusere en 20-årsflom til en 10-årsflom i en time trengs 26640 m3, for ett døgn ca 640000 m3. Tilsvarende kan en redusere en 100-årsflom til en 50-årsflom i en time med 35000 m3. Dette viser at arealet har potensiale for å dempe flomtoppene.....

Alleen beschikbaar als e-document

Alleen beschikbaar als e-document

Lit.opg., bijl.

Lit.opg., bijl.

Met lit. opg.