Sensitivity-guided reduction of parametric dimensionality for multi-objective calibration of watershed models

Problem complexity for watershed model calibration is heavily dependent on the number of parameters that can be identified during model calibration. This study investigates the use of global sensitivity analysis as a screening tool to reduce the parametric dimensionality of multi-objective hydrological model calibration problems while maximizing the information extracted from hydrological response data. This study shows that by expanding calibration problem formulations beyond traditional, statistical error metrics to also include metrics that capture indices or signatures of hydrological function, it is possible to reduce the complexity of calibration while maintaining high quality model predictions. The sensitivity-guided calibration is demonstrated using the Sacramento Soil Moisture Accounting (SAC-SMA) conceptual rainfall–runoff model of moderate complexity (i.e., up to 14 freely varying parameters). Using both statistical and hydrological metrics, optimization results demonstrate that parameters controlling at least 20% of the model output variance (through individual effects and interactions) should be included in the calibration process. This threshold generally yields 30–40% reductions in the number of SAC-SMA parameters requiring calibration – setting the others to a priori values – while maintaining high quality predictions. Two parameters are recommended to be calibrated in all cases (percent impervious area and lower zone tension water storage), three parameters are needed in drier watersheds (additional impervious area, riparian zone vegetation, and percent of percolation going to tension storage), and the lower zone parameters are crucial unless the watershed is very dry. Overall, this study demonstrates that a coupled, multi-objective sensitivity and calibration analysis better captures differences between watersheds during model calibration and serves to maximize the value of available watershed response time series. These contributions are particularly important given the ongoing development of more complex integrated models, which will require new tools to address the growing discrepancy between the information content of hydrological data and the number of model parameters that have to be estimated.     

Lightning evolution related to radar-derived microphysics in the 21 July 1998 EULINOX supercell storm

Results of a combined analysis of data from a C-band polarimetric Doppler radar and a 3D VHF interferometric lightning mapping system, as obtained during the European Lightning Nitrogen Oxides project (EULINOX) field campaign, are presented. For 21 July 1998, the lightning data from a supercell thunderstorm weakly indicate a tendency for a bi-level vertical distribution of lightning VHF emissions around the −15°C and −30°C temperature levels. Also, in some parts of the clouds, evidence is found for the presence of a lower positive charge center near the freezing level. However, where strong vertical motions prevail, VHF emissions are not organized in horizontal layers but in oblique or vertical regions. Correlation of VHF signals with radar quantities shows that in the growing storm, peak VHF activity is low and related to reflectivity factors around 30 dBZ, while after the mature stage, the peak VHF activity is about three times larger. The highest density of VHF signals is now found near reflectivity factors of 45 dBZ. A polarimetric hydrometeor classification indicates that during storm development, most lightning activity occurs where graupel and, secondarily, snow and small dry hail are present. In the decaying phase of the supercell hailstorm, however, most lightning VHF emissions stem from the region with hail and heavy rain. Furthermore, while the VHF signal frequency per cubic kilometer in the graupel and rain regions remains nearly constant throughout the supercell life cycle, the signal frequency in the hail region rises during storm decay.     

Mesoscale Variability in the Summer Arctic Boundary Layer

Observations from the summer Arctic Ocean Experiment 2001 (AOE-2001) are analysed with a focus on the interactions between mesoscale and boundary-layer dynamics. Wavelet analyses of surface-pressure variations show daylong periods with different characteristics, some featuring episodes of pronounced high-frequency surface-pressure variability, here hypothesized to be caused by trapped gravity waves. These episodes are accompanied by enhanced boundary-layer turbulence and an enhanced spectral gap, but with only minor influence on the surface stress. During these episodes, mesoscale phenomena were often encountered and usually identified as front-like features in the boundary layer, with a peak in drizzle followed by changing temperature. These phenomena resemble synoptic fronts, though they are generally shallow, shorter-lasting, have no signs of frontal clouds, and do not imply a change in air mass. Based on this analysis, we hypothesize that the root cause of the episodes with high-frequency surface-pressure variance are shallow, mesoscale fronts moving across the pack ice. They may be formed due to local-to-regional horizontal contrasts, for example, between air with different lifetimes over the Arctic or with perturbations in the cloud field causing differential cooling of the boundary layer. Thermal contrasts sharpen as the air is transported with the mean flow. The propagating mesoscale fronts excite gravity waves, which affect the boundary-layer turbulence and also seem to favour entrainment of free tropospheric air into the boundary layer.     

Garnet‐controlled very low velocities in the lower mantle transition zone at sites of mantle upwelling

Deep mantle plumes and associated increased geotherms are expected to cause an upward deflection of the lower–upper mantle boundary and an overall thinning of the mantle transition zone between about 410 and 660 km depth. We use subsequent forward modelling of mineral assemblages, seismic velocities, and receiver functions to explain the common paucity of such observations in receiver function data. In the lower mantle transition zone, large horizontal differences in seismic velocities may result from temperature‐dependent assemblage variations. At this depth, primitive mantle compositions are dominated by majoritic garnet at high temperatures. Associated seismic velocities are expected to be much lower than for ringwoodite‐rich assemblages at undisturbed thermal conditions. Neglecting this ultralow‐velocity zone at upwelling sites can cause a miscalculation of the lower–upper mantle boundary on the order of 20 km.     

Regime-dependent evaluation of accumulated precipitation in COSMO

Regime-dependent evaluation is a relatively new approach to assess model performance. It consists of classifying the model biases according to a discrete number of regimes and evaluating model output within each regime. In this paper, the regimes are firstly defined by the large-scale atmospheric circulation, based on the objective Jenkinson-Collison classification technique which distinguishes synoptic patterns by strength, direction and vorticity of the geostrophic flow. Eight directional and two vorticity circulation regimes (circulation types) are specified. In this way, it is possible to quantify the model performance for cases with for example westerly winds only, or with cyclonic circulation only. A second regime classification is based on temperature, which allows for detection of temperature-dependent model performance. Modelled accumulated precipitation (mm/6?h) is evaluated with rain gauges for the years 2007 and 2008. Two variants of the COSMO model are evaluated: a fine-resolution version (2.8?km, COSMO-DE) and a coarse-resolution version (7?km, COSMO-EU). In COSMO-EU, a windward/leeward effect becomes visible since circulation is related to dominant wind direction, hence to windward and lee side of orography. In COSMO-DE, no circulation dependent but a height-related bias is identified and further explored, making use of temperature-dependent evaluation which unveils a positive model bias related to solid precipitation.     

A biomarker and δ15N study of thermally altered Silurian cyanobacterial mats

Early Silurian cherts from the Holy Cross and Bardzkie Mountains (Poland) contain abundant microfossils morphologically resembling contemporary cyanobacteria. Most of the organic matter preserved in the cherts is highly mature and extensively degraded because of biological decomposition and progressive thermal alteration. These processes may have changed the original morphology of the deposited microbial remains, so the microfossil origin could be easily misinterpreted. The cherts were therefore examined using organic geochemical and stable isotope techniques to provide support for the presence of cyanobacterial remains. The nitrogen isotopic composition of bulk sediments and extractable organic matter ranged from +0.1‰ to ?2.2‰ and from +1.8‰ to ?1.7‰, respectively. The δ¹⁵N values are thus in good agreement with a contribution of diazotrophic cyanobacteria for both locations. Biomarkers in the Holy Cross Mts. cherts included mid-chain branched monomethyl alkanes, indicative of a cyanobacterial contribution. However, molecular fossils of a cyanobacterial origin were not detected in the Bardzkie Mts. cherts, most likely because of the greater maturity than those from the Holy Cross Mts.     

Hf-W chronology of the accretion and early evolution of asteroids and terrestrial planets

The ¹⁸²Hf-¹⁸²W systematics of meteoritic and planetary samples provide firm constraints on the chronology of the accretion and earliest evolution of asteroids and terrestrial planets and lead to the following succession and duration of events in the earliest solar system. Formation of Ca,Al-rich inclusions (CAIs) at 4568.3 ± 0.7 Ma was followed by the accretion and differentiation of the parent bodies of some magmatic iron meteorites within less than ∼1 Myr. Chondrules from H chondrites formed 1.7 ± 0.7 Myr after CAIs, about contemporaneously with chondrules from L and LL chondrites as shown by their ²⁶Al-²⁶Mg ages. Some magmatism on the parent bodies of angrites, eucrites, and mesosiderites started as soon as ∼3 Myr after CAI formation and may have continued until ∼10 Myr. A similar timescale is obtained for the high-temperature metamorphic evolution of the H chondrite parent body. Thermal modeling combined with these age constraints reveals that the different thermal histories of meteorite parent bodies primarily reflect their initial abundance of ²⁶Al, which is determined by their accretion age. Impact-related processes were important in the subsequent evolution of asteroids but do not appear to have induced large-scale melting. For instance, Hf-W ages for eucrite metals postdate CAI formation by ∼20 Myr and may reflect impact-triggered thermal metamorphism in the crust of the eucrite parent body. Likewise, the Hf-W systematics of some non-magmatic iron meteorites were modified by impact-related processes but the timing of this event(s) remains poorly constrained.The strong fractionation of lithophile Hf from siderophile W during core formation makes the Hf-W system an ideal chronometer for this major differentiation event. However, for larger planets such as the terrestrial planets the calculated Hf-W ages are particularly sensitive to the occurrence of large impacts, the degree to which impactor cores re-equilibrated with the target mantle during large collisions, and changes in the metal-silicate partition coefficients of W due to changing fO₂ in differentiating planetary bodies. Calculated core formation ages for Mars range from 0 to 20 Myr after CAI formation and currently cannot distinguish between scenarios where Mars formed by runaway growth and where its formation was more protracted. Tungsten model ages for core formation in Earth range from ∼30 Myr to >100 Myr after CAIs and hence do not provide a unique age for the formation of Earth. However, the identical ¹⁸²W/¹⁸⁴W ratios of the lunar and terrestrial mantles provide powerful evidence that the Moon-forming giant impact and the final stage of Earth’s core formation occurred after extinction of ¹⁸²Hf (i.e., more than ∼50 Myr after CAIs), unless the Hf/W ratios of the bulk silicate Moon and Earth are identical to within less than ∼10%. Furthermore, the identical ¹⁸²W/¹⁸⁴W of the lunar and terrestrial mantles is difficult to explain unless either the Moon consists predominantly of terrestrial material or the W in the proto-lunar magma disk isotopically equilibrated with the Earth’s mantle.Hafnium-tungsten chronometry also provides constraints on the duration of magma ocean solidification in terrestrial planets. Variations in the ¹⁸²W/¹⁸⁴W ratios of martian meteorites reflect an early differentiation of the martian mantle during the effective lifetime of ¹⁸²Hf. In contrast, no ¹⁸²W variations exist in the lunar mantle, demonstrating magma ocean solidification later than ∼60 Myr, in agreement with ¹⁴⁷Sm-¹⁴³Nd ages for ferroan anorthosites. The Moon-forming giant impact most likely erased any evidence of a prior differentiation of Earth’s mantle, consistent with a ¹⁴⁶Sm-¹⁴²Nd age of 50-200 Myr for the earliest differentiation of Earth’s mantle. However, the Hf-W chronology of the formation of Earth’s core and the Moon-forming impact is difficult to reconcile with the preservation of ¹⁴⁶Sm-¹⁴²Nd evidence for an early (<30 Myr after CAIs) differentiation of a chondritic Earth’s mantle. Instead, the combined ¹⁸²W-¹⁴²Nd evidence suggests that bulk Earth may have superchondritic Sm/Nd and Hf/W ratios, in which case formation of its core must have terminated more than ∼42 Myr after formation of CAIs, consistent with the Hf-W age for the formation of the Moon.     

Analysis of the Moberg et al. (2005) hemispheric temperature reconstruction

The Moberg et al. (Nature 433(7026):613–617, 2005. doi:10.1038/nature03265; M05) reconstruction of northern hemisphere temperature variations from proxy data has been criticised; the M05 method may artificially inflate low-frequency variance relative to reality. We test this assertion by undertaking several pseudoproxy experiments in three climate model simulations—one control run and two forced simulations that include several time-varying radiative forcings. The pseudoproxy series are designed to have the same variance spectra as the real M05 proxies, primarily to mimic the low-resolution character of several series. A simple composite-plus-scale (CPS) method is also analysed. In the CPS case all input data behave like annually resolved proxies. The spectral domain performance of both M05 and CPS is found to be dependent on the noise type and noise level in pseudoproxies, on the variance spectrum of the climate model simulation, and on the degree of data smoothing. CPS performs better than M05 in most investigated cases with the control run, but leads to deflated low-frequency variance in some cases. With M05, low-frequency variance tend to be inflated for the control run but not for one of the forced runs and only very slightly with the other forced simulation. Hence, the M05 approach does not routinely inflate low-frequency variance. In our experiment, the M05 approach performs better in the spectral domain than CPS when applied to forced climate model simulations. The results underscore the importance of evaluating the variance spectrum of climate reconstructions.