In past global dust storms, no long lasting anomalies in the pressure cycle had been observed. The Global Dust Storm of Mars Year 34 (MY34), however, left behind an average surface pressure lower than what was expected based on the the values recorded on previous years by the Rover Environmental Monitoring Station (REMS) on Curiosity. The main signal contribution to the daily average surface pressure is the CO2 cycle, which is controlled by the Polar ice sublimation and freezing cycles. We used REMS and Mars Climate Sounder (MCS) data to search for correlations between the REMS anomaly and anomalies in the circulation compared to MCS observations from previous years. The findings include an early start of the retreat season for the Northern Polar cap, followed by the longest period of growth for the Southern Polar (SP) cap ice expansion since Curiosity had landed and then, during the dust storm, the longest retreat season of the Southern Polar cap. We also find a larger Northern Polar Cap extension after the storm, suggestive of a larger deposition of CO2 ice. The changes in length of the SP growth and retreat seasons might be consequence of the response of the zonal mean circulation to the dust storm. Changes in the structure of the zonal mean circulation compared to previous years are found in MCS data and presented. The combination of these anomalies constraint what physical processes may have caused this response in surface pressure after the dust storm.

Deformation bands are the main structural element of fault damage zones within sandstone reservoirs. The prediction of band occurrence and their petrophysical impacts is based largely on the understanding that the yield and deformation mechanism of sandstones is primarily controlled by porosity and mean grain size. Whilst this is supported by field observations within aeolian successions, where bands are predictably favoured within coarse-grained, high-porosity sandstones, the prediction of deformation bands within texturally complex mixed aeolian-fluvial reservoirs on the basis of porosity and grain size alone, may be unreliable. The effect of grain sorting on the mechanical behaviour of sandstones is not well understood, although it is generally regarded that deformation band formation is inhibited in texturally immature sandstones with a poor level of sorting. We examine the effect of sorting on both the inelastic yield of sandstones, the dominant deformation mechanism by which yield occurs, and the textural and microstructural changes with deformation, using a series of triaxial experiments on unconsolidated quartz sands. Hydrostatic experiments were conducted on over-consolidated samples of very well- to moderately-sorted sands with a range of mean grain sizes from 128-700µm. We report accurate prediction of P* using porosity x grain radius, with P* reduced with decreased sorting. Constant displacement rate triaxial experiments are performed at up to 10% axial strain to explore yield behaviour in both the brittle dilatant regime and shear-enhanced compactive regime. Experiments were repeated with systematically varied grain sorting whilst mean grain size and porosity was maintained. The textural and petrophysical changes are observed and quantified using pore volumometry, back scattered electron microscopy, digital image analysis and point counting. Results show that in well-sorted sands, localised cataclasis and deformation band formation is the dominant deformation mechanism. In poorly-sorted sands deformation occurs through a combination of grain boundary sliding and randomly distributed pockets of cataclasis. Using grain size analysis we identify greater levels of cataclasis and production of fines in well-sorted sands, resulting in permeability reduction up to one order of magnitude more than that of poorly-sorted sands deformed at the same conditions. We hypothesise that band formation within poorly sorted sandstones may be promoted by the formation and propagation of bands in adjacent well sorted sandstones where band formation is favoured. These results give insight into the deformation, textural changes, and permeability impact of both unconsolidated and consolidated siliciclastic reservoirs.

Soft x-ray and EUV radiation from the Sun is absorbed by and ionizes the atmosphere, creating both the ionosphere and thermosphere. Temporal changes in irradiance energy and spectral distribution can have drastic impacts on the ionosphere, impacting technologies such as satellite drag and radio communication. Because of this, it is necessary to estimate and predict changes in Solar EUV spectral irradiance. Ideally, this would be done by direct measurement but the high cost of solar EUV spectrographs makes this prohibitively expensive. Instead, scientists must use data driven models to predict the solar spectrum for a given irradiance measurement. In this study, we further develop the Synthetic Reference Spectral Irradiance Model (SynRef). The SynRef model, which uses broadband EUV irradiance data from EUVM at Mars, was created to mirror the SORCE XPS model which uses data from the TIMED SEE instrument and the SORCE XPS instrument at Earth. Both models superpose theoretical Active Region and Quiet Sun spectra generated by CHIANTI to match daily measured irradiance data, and output a modeled solar EUV spectrum for that day. By adjusting the weighting of Active Region and Quiet Sun spectra, we update the SynRef model to better agree with the FISM model and with spectral data collected from sounding rocket flights. We also use the broadband EUVM measurements to estimate AR temperature. This will allow us to select from a library of AR reference spectra with different temperatures. We present this updated SynRef model to more accurately characterize the Solar EUV and soft x-ray spectra.

The unprecedented growth of emissions has deteriorated air quality dramatically leading to a pulmonary complication on human health especially during winter season increasing the prevalence of Chronic Obstructive Pulmonary Diseases (COPD) more in female compared to male. Selecting different peak and non-peak hour this study estimated vehicular emission load with the help of emission factors, derived equations, field visits, and literature review. The average annual vehicular energy demand of Bhaktapur municipality was estimated 33,044 GJ. The emission load of Bhaktapur municipality was estimated 3,310 tons/year including (CO2, CO, NOx, HC and PM10) of which CO2 accounts for 94.36% of total emissions followed by CO(4.39%), HC(0.72%), NOx (0.35%), and PM10 (0.18%) respectively. Statistical analysis showed significant positive correlation (r = 0.92, p = 0.002) between CO2 and PM10, (r = 0.87, p = 0.009) between CO2 and NOx, (r = 0.90, p = 0.004) between CO and HC, (r = 0.74, p = 0.05) between NOx and PM10 respectively. Assuming an inauguration of electric Car, Motorbike and Buses within the Municipality at the rate of 10%, 20%, and 30% showed a significant reduction in emissions by 157 tons/year, 314 tons/year and 471 tons/year respectively. Carbon dioxide was found more potent to deteriorating air quality in the future compared to other vehicular pollutants. Replacement of fossil fuel based technology with clean energy technology is needed in the Bhaktapur municipality to sustain human health and welfare. Having lower emissions in Bhaktapur Municipality compared to its nearest city Kathmandu despite differences in population density and area size, effective preparedness, policy, and response in the present scenario can prevent worse possible future scenario.

Automated classification of remote sensing data is an integral tool for earth scientists, and deep learning has proven very successful at solving such problems. However, building deep learning models to process the data requires expert knowledge of machine learning. We introduce DELTA, a software toolkit to bridge this technical gap and make deep learning easily accessible to earth scientists. Visual feature engineering is a critical part of the machine learning lifecycle, and hence is a key area that will be automated by DELTA. Hand-engineered features can perform well, but require a cross functional team with expertise in both machine learning and the specific problem domain, which is costly in both researcher time and labor. The problem is more acute with multispectral satellite imagery, which requires considerable computational resources to process. In order to automate the feature learning process, a neural architecture search samples the space of asymmetric and symmetric autoencoders using evolutionary algorithms. Since denoising autoencoders have been shown to perform well for feature learning, the autoencoders are trained on various levels of noise and the features generated by the best performing autoencoders evaluated according to their performance on image classification tasks. The resulting features are demonstrated to be effective for Landsat-8 flood mapping, as well as benchmark datasets CIFAR10 and SVHN.

Ground-based microwave radiometry is a common tool to estimate profiles of the atmosphere. With a high temporal resolution radiometers have became an alternative to atmospheric sounding like radiosondes. However remote sensing radiometry requires the use of inversion algorithms, where methods like linear-, quadratic-regression or Artificial Neural Network are commonly used. The present study implements a Bayesian inversion technique as alternative to the state-of-the-art retrieval algorithms provided by the radiometer’s manufacturer firmware. The Bayesian inversion provides advantages over other established methods, namely: the use of a-priori suited for the specific climatology under observation, the estimation of the most likely profile along with its uncertainty obtained from the posteriori distribution, and the feasibility to add synergistic observations from other instruments to increase retrieval capabilities. To estimate the uncertainties resulting from the Bayesian and firmware retrieval algorithms, synthetic radiometer data have been created by means of radiative transfer simulations using radiosonde profiles as descriptor of atmospheric states. These synthetic data mimics the instrument’s firmware binary files letting the radiometer to perform retrievals as real measurements. By analyzing the differences from retrieval results relative to the known true profile we assess uncertainty metrics to characterize the algorithms. It has been found that Bayesian inversion reproduces more accurately the profile vertical structure as compared to the firmware, specially for humidity profiles. Absolute errors have been strongly reduced mainly at the lower atmosphere. The study concludes that Bayesian inversion for ground-based atmospheric profiling produces results resembling observations by radiosondes when a suitable a-priori distribution is used.

The agriculture sector consumes more than two-thirds of world’s limited freshwater resources. However, only a small part of the water (less than 5%) that is taken up by roots is used for plant growth, while the rest (above 95%) is lost due to transpiration through the stomatal apertures. Therefore, reducing the transpiration of agricultural plants will contribute to the preservation of precious water resources. However, reducing the transpiration rate artificially is difficult because most plants react delicately and negatively, resulting in water-stressed conditions that often cause different physiological disorders. The present study investigated the transpiration light response in tomato plants (Solanum lycopersicum) grown under LED lights and assessed different irradiation techniques’ ability to reduce transpiration and maintained proper plant growth in a controlled environment. Tomato plants were grown in three enclosed hydroponic units under blue (460 nm) and red (630 nm) LEDs inside an air-conditioned glasshouse. The test plants and multiple replicates were grown five consecutive times, and the irradiation intensity (photosynthetic photon flux density (PPFD)), irradiation pattern (simultaneous/alternate irradiation for red/blue LEDs) and LED combination (number/ratio of red/blue LEDs) were changed each time. The plants’ physiological parameters (transpiration, stomatal conductance, stem-diameter, stem height, and number of leaves) and daily transpiration rates were recorded periodically and analyzed. The results show that a typical photoperiod of 12 hours with simultaneous irradiation of red/blue LEDs produced balanced physiological growth for plants in general. However, when normalized against water use efficiency (transpiration), an alternate irradiation pattern (6 hours: blue LED on/off repeatedly for 15-minute intervals + 6 hours: red LED on/off repeatedly for 15-minute intervals) was the most suitable for tomato cultivation in controlled environments.

It has become apparent in recent years that scientists need to find new ways to communicate and connect with the public to increase science literacy and trust of scientific results. To address these issues, the Time Scavengers website (timescavengers.blog) was created. This website is maintained and continuously added to by a team of collaborators including graduate students, post docs, museum staff, professors, avocational scientists, educators, and an editor. The website also includes static pages on the scientific method, geology, and climate science methods, as well as a number of resources for educators and others interested in science. The collaborators contribute regular blog posts on a variety of topics related to being a scientist, including the work we do in the field, learning new methods, and various aspects of our academic and career paths. One of our more popular blogs is called ‘Meet the Scientist’, which showcases diverse scientists in many different fields, from graduate students to experienced professional scientists, both U.S.-based and international. The website has reached almost 63,000 unique visitors in the two years since it was created, reaching folks speaking 155 languages in 196 countries. Using data from Google Analytics and social media accounts, including Facebook, Twitter, and Instagram, we examined some of the trends related to our broad international reach, to determine if any specific posts or types of posts attracted more international or non-English speaking visitors. Besides examining the general geographic reach over time, a few more specific comparisons were conducted. We examined whether or not Meet the Scientist posts featuring international scientists attracted more international visitors than those featuring U.S.-based scientists. We also analyzed data forField Excursions posts that described places people could visit to see if they attracted site visitors from those areas described in the post or had a broader national and international reach. Preliminary data indicate that posts about international scientists reach more countries, on average, than those featuring U.S. scientists, and geographic-specific posts reach a broad national and international audience.

The quantification and monitoring of photosynthesis are essential to understand the global carbon cycle and vegetation’s responses to climate. Among the different remotely-sensed photosynthesis-related variables, Sun-Induced chlorophyll a Fluorescence (SIF) is especially promising since it results directly from photochemical energy conversion but uncertainties still complicate its interpretation. Recent studies have pointed to the influences of vegetation biochemistry and structure on radiative transfer as the main confounding factors for the use of SIF as a photosynthesis proxy. Leaf-level fluorescence research has shown that such influences may be removed by adjusting the raw fluorescence signal to the emitting leaf’s spectra and we suggest that this can be upscaled to the landscape level. In this study we present and test new Spectrally-Adjusted SIF formulations (SASIFs), along with previously proposed SIF modifications and other acknowledged photosynthesis productivity proxies, against carbon-flux data from vegetation of diverse structure. Accordingly, we used Gross Primary Productivity (GPP) data spanning periods from two to seven years, from 27 FLUXNET sites classified into different Land Cover Classes (LCCs) as defined by the International Geosphere-Biosphere Programme (IGBP). The data tested against GPP was calculated with GOME-2 SIF data, MODIS reflectance and spectral vegetation indices, and it included: NIRV, SIF from the red and the far-red frequency peaks, SIF normalized by the cosine of the Sun’s zenith angle, SIF-yield, new SASIFs and FLUXCOM GPP. The relationships between all variables and FLUXNET GPP were tested using time-series decomposition, site- and LCC-specific Kendall’s rank correlation tests and linear mixed model analysis. Results show that one of our new SASIFs has the best overall correlation to FLUXNET GPP among all tested data. Our LCC-specific analysis demonstrates the influences of biochemistry, phenology, temporal resolution and vegetation structure on the relationships between the tested variables. Results support the idea that chlorophyll fluorescence can be complemented with reflectance data improving our ability to monitor vegetation productivity and predict climate-driven changes to standing biomass in spite of their particular limitations.

Bangladesh, a small and over populated country in Southeast Asia occupies most of the Bengal Basin that results from sediments derived from the collision of India with Asia. The basin is filled with a 19 km thick sequence of Cenozoic sediments deposited by the mighty rivers Ganges and Brahmaputra. Unconsolidated Holocene sediments susceptible to seismic amplification characterize the upper part of the Cenozoic sequence. Bangladesh sits a top on three tectonic plates; India, Tibet and Burma. The India plate is colliding with the Tibet subplate to the north, which gives rise to great Himalayas, while to the east it is subducting beneath Burma and Sunda slivers, which gave rise to Indo-Burma arc. The Surma basin of NE Bangladesh is being underthrust under the Shillong massif producing the 2-km high plateau. The Indo-Burma fold and thrust belt results from the oblique subduction of the thick sediments of the Bengal Basin on the India plate that has deformed into a series of north-south trending en-echelon folds and thrust faults. The faults rooting these folds and the underlying megathrust are capable of generating devastating earthquakes in and around Bangladesh. Past earthquakes have brought changes to the landscape, avulsion of rivers Brahmaputra and Meghna, migration of human settlements, and widespread sand liquefactions and sand and/or mud eruptions. Our GPS study demonstrated that the landward extension of Andaman-Sumatra subduction zone into Indo-Burma subduction in deltaic Bangladesh is active. The present day India-Burma oblique convergence rate is 17 mm/y and that the décollement beneath the fold-thrust belt is locked (Steckler et. al., 2016). The western part of the subduction zone over a shallow décollement shows little seismicity whereas the eastern part shows moderate seismicity of magnitude 4 to 6. Based on the GPS velocity across the fold belt and seismicity the Indo-Burma subduction zone can be potentially be divided into locked western segment and slipping eastern segment, analogous to Cascadia subduction zone. Fold belt parallel shortening across Dauki Fault in Shillong is 7 mm/yr, which is another potential source of a large earthquake. The huge population might be severely ravaged by devastating earthquakes from both these sources.

The 2016 Amatrice-Norcia seismic sequence in central Italy activated a system of normal faults in the central Apennines and ruptured the surface along the Monte Vettore normal fault. Due to the complex rupture behavior, including antithetic faults and the proposed reactivation of an old thrust front, the Amatrice-Norcia seismic sequence offers a unique opportunity to study the relationship between fault complexity, surface ruptures, and earthquake source properties. Here, we focus on the first two months of the Amatrice-Norcia seismic sequence, including the 30 October 2016 Mw 6.5 mainshock near Norcia and more than 25000 aftershocks. Using continuous waveform data from 94 seismic stations with epicentral distances of up to ~100 km, we estimate source parameters of all cataloged earthquakes that exceed specific quality control criteria in a time period ranging from 24 October – 29 November 2016. Displacement spectral corner frequency and seismic moment values are fit using individual earthquake spectra, and corner frequency estimates are refined using spectral ratios. Constrained spectral parameters then provide input for static stress drop estimates based on a circular crack model. Preliminary results suggest the majority of earthquakes have static stress drop values between 1 and 10 MPa and self-similar scaling. Due to the high quality and quantity of available data, including precise earthquake locations, manually reviewed phase arrivals, and detailed mapping of surface ruptures, the Amatrice-Norcia earthquake sequence represents an opportunity to link earthquake source parameters to geological structures and surface rupture complexity. Preliminary results show correlations between high stress drop values and areas with increasing fault complexity, such as fault intersections at depth (inferred from precise earthquake hypocenters) or the mapped tip of the Monte Vettore normal fault, relative to other fault patches with fewer intersections or mapped surface trace terminations. Future work will examine whether the correlation of stress drop and fault complexity holds using refined stress drop estimates obtained using spectral ratio approaches.

This paper presents microphysical inference retrievals obtained from spectral polarimetry during the Relampago (Remote sensing of Electrification, Lightning, And Mesoscale/Microscale Processes with Adaptive Ground Observations) campaign. Spectral processing has been an essential part of weather radar moments estimation for a long period of time. Various processing can be performed in the spectral domain including precipitation detection in presence of strong clutter and noise, clutter & interference mitigation by algorithms such as GMAP, object-oriented filters and many more. However spectral applications to polarimetry have been rare. The C band CSU-CHIVO radar that was deployed in Cordoba region in Argentina between June 2018 and April 2019 during the Relampago campaign, recorded some of the tallest storms in the world characterized by strong wind shear, updrifts, turbulence and occurrence of severe hail and rain. The polarimetric spectrum in precipitation with rain and hail mixtures were characterized. This Spectral polarimetry revealed different spectral characteristics including multi-modal spectrum, spectral broadening, slopes in spectral differential reflectivity and lowering of coherency spectrum. These results characterized occurrence of mixed hydrometeor types in a radar resolution volume such as presence of rain and hail mixture, large drops formation and size sorting. Spectral displays are inherently noisy, and the paper also presented methodology to obtain clean quality spectrum implementing spectral quality index, that is used to process the observations and the results are presented.

Learning progressions provide a sequence, or progression, of concepts from naive to sophisticated. Astrobiology educators and scientists have identified the need to develop learning progressions for core, interdisciplinary concepts in astrobiology to support both educators of K-12 students to bring astrobiology concepts into their classrooms, and scientists to communicate with a range of audiences. The Astrobiology Learning Progressions resource organizes core concepts around the essential questions of astrobiology, and includes connections to the Next Generation Science Standards, progressed storylines, and concept boundaries for four levels: primary or adult naïve learners, elementary or emerging adult learners, middle school or building learner, and high school or sophisticated learner. The resource also links lesson plans and other learning materials to each core concept.

The Hudson Bay Lowlands (HBL) is a vast continuous peatland in Northern Canada. The landscape is a mosaic of mostly bogs and fens, with more limited swamp, marsh, forest and open water. Owing to rapid rates of isostatic uplift, younger peats are found closer to the coasts of Hudson and James Bays, with fen-type peatlands somewhat more prevalent on these younger surfaces. More than 30 Pg of carbon have accumulated in the HBL over the Holocene. The rates of Holocene carbon accumulation vary considerably both spatially and temporally, with some sites showing more rapid rates of carbon accumulation in the first 2-3 millennia following peatland initiation. We evaluate here the hypothesis that vegetation changes over the course of the Holocene, including fen-to-bog transitions, partially explain the variability in carbon accumulation. We find that in some cases, more rapid rates of C accumulation in the middle Holocene (5000-8000 yrs before present) are associated with early successional minerotrophic fens with higher carbon densities. Fen-to-bog transitions are recorded in many peat cores collected from present day bogs; however, these transitions are time transgressive, and can depend on the time since initiation, suggesting that climate changes may play a secondary role, relative to hydrological changes and local ecological processes. Fens are highly prevalent in the HBL landscape (covering about 38% of land cover). Cores taken from present day fens and analyzed for carbon accumulation and vegetation change indicate that many fen sites have remained fens since peat initiation. Variability in rates of Holocene carbon accumulation within fen records which have not been subject to any major vegetation change may more closely reflect climate drivers.

Validation of cloud hydrometeors simulation from the global models is important issue as it pertains to the accuracy of climate predictions. In this study, the cloud hydrometeor data from Korean Integrated Model (KIM) is validated using different Reanalysis (ERAI, ERA5, and MERRA) and Satellite Observations (Cloudsat). In ERA5 products, cloud snow water and rain water are also available. Satellite observations are gridded to compare with model simulations. Cloud liquid water (Qc), Cloud snow water (Qs), Cloud ice water (Qi), Cloud rain water (Qr), Vapour mixing ratio (Qv) for January (dry) and July (wet seasons) of 2017 are considered for validation. BIAS and RMSE are calculated for comparison. To understand the vertical distribution of hydrometeors, contour frequency altitude diagrams (CFADs) are plotted. Early validation of KIM hydrometeors shows the reasonable estimate of different hydrometeors with KIM model showing more Qc at surface, more Qv at upper levels. The vertical structure of Qi has showed significant bias at upper levels with model showing large ice values at higher levels. ERAI and ERA5 products has showed distinct pattern of Qi due to different configurations. More Qs at upper levels is also evident in model simulations. Combined distribution (Qc+Qi) of KIM at lower (upper) levels is more comparable with ERA5 (MERRA) products. Further, Qr distribution shows underestimation at the equator and over estimation at the latitude belts. To examine the contribution of different physics modules related to the bias, the hydrometeors from cumulus, microphysics and shallow convection are also analyzed separately. Accuracy of KIM simulated cloud hydrometeors against different products and possible causes for biases will be discussed in the conference.

Western U.S. conifer forests harbor diverse ecological strategies that enable species to persist across a wide range of hydroclimate conditions, along with wildfire and eruptive insect outbreaks. Assessing climate influences on future forest composition and carbon sequestration requires vegetation process models that have sufficient ecological resolution to simulate this range of ecological variability. Here we present progress towards incorporating multiple shade and drought tolerance strategies in a vegetation demographic model parameterized for Western U.S. forests. We used the Functionally Assembled Terrestrial Ecosystem Simulator (FATES) to simulate a mixed conifer forest dominated by ponderosa pine and incense cedar in the Sierra Nevada Mountains of California. FATES resolves plant growth and respiration at the level of cohorts, defined by size and plant functional type. Incense cedar is shade and drought tolerant, while ponderosa pine is shade intolerant and the canopy dominant. We synthesized literature values of plant traits that correspond to important physiological and allometric parameters in FATES and conducted a sensitivity analysis within the observed parameter ranges with respect to carbon and water fluxes. Model output was benchmarked against carbon flux, water flux, and leaf area index measurements from the Critical Zone Observatory/AmeriFlux CZ2 site during 2010-2012. Specific leaf area, Vcmax, rooting distribution, and allometric equations had the most influence on simulated carbon and water fluxes. Final simulated average annual gross primary production (GPP) over 2010-2012 (1156 +- 79.2 gC/m2/yr) was 3.8% lower than observed GPP (1202 +-138.2 gC/m2/yr). Simulated evapotranspiration (ET, 373 +- 25 mm/yr) was 62% lower than measured ET (993 +-158 mm/yr). Simulated leaf area index (LAI, 1.2) was within the range of measured LAI (0.5-1.5). Preliminary analysis indicates underestimation of ET is likely due to an overestimation of soil water drainage. Our final parameter set allows pine and cedar coexistence to emerge from a bare ground initialization, and additional sensitivity testing of parameters important for coexistence are in progress. Clearly, observationally constrained parameters are critical for simulating ecosystem dynamics in Western U.S. forests.

Deep-sea δ18O records show a pronounced difference in Milankovitch periodicity between the Early and Late Pleistocene. δ18O is interpreted as a proxy for ice sheet volume and temperature, which led to the conclusion that glacial-interglacial cycles considerably changed their rhythm during the mid-Pleistocene. This transition is referred to as the mid-Pleistocene Transition (MPT). Specifically, the precessional component of the Milankovitch cycles is absent in Early Pleistocene δ18O records, despite its continuous presence in solar insolation forcing to the ice sheets. Climate feedbacks involving (sea) ice, geological processes and carbon and nutrient cycling have been proposed as causes of this marked change. We however show that the absence of an Early Pleistocene precession signal in deep-sea δ18O records could be the result of destructive interference of the precessional cycle in the interior ocean. Such cancellation is caused by the anti-phasing of the precessional cycle between the North Atlantic and Southern Ocean deep-water sources (see Figure). We explore the potential for cancellation in the transient setup of the Total Matrix Intercomparison model for a wide range of source signal strengths. Our results show that cancellation can cause the absence of the precessional signal due to cancellation in the interior South-Atlantic, Indian and Pacific basins. Cancellation is especially widespread for a relative end-member contribution typical for the Early Pleistocene. We therefore conclude that the precessional component is likely incompletely archived in Early Pleistocene δ18O records, and appears as an actual change in Milankovitch periodicity across the MPT. Proxies not susceptible to cancellation of precession (such as those currently retrieved across the MPT from Antarctica) would be able to verify to what extent deep-sea δ18O correctly represents Pleistocene climate.

The watershed determined by Aburrá Valley system, located in northwestern Colombia, has significant urban development and steep hills. These features, together with the typical intense storms of the region, make the watershed prone to the occurrence of flash floods during the rainy seasons, affecting vulnerable communities. We propose a hybrid observational-modeling strategy to generate 30-minute discharge forecasts in different locations of the watershed, using an operational distributed hydrological model, information from stream gauges, and weather radar-derived precipitation using a quantitative precipitation estimation (QPE) technique. The forecast methodology is triggered when any stream gauge of interest reports levels over a predefined threshold. As a first step, the model uses different rainfall scenarios for the following 30 minutes. Every 5 minutes, the model forecast is executed after updating the observed rainfall and the rainfall scenarios. The scenarios correspond to (i) a lagrangian extrapolation of the precipitation fields, (ii) to a cellular automata-based extrapolation and to (iii) the last observed rain field multiplied by a time-varying ad-hoc factor based on historical event analysis. To parametrize the hydrological model and to validate the prediction methodology, we use 173 storm events from 2013 to 2018. The methodology is evaluated using the Nash coefficient, the Klin-Gupta index, differences in time-to-peak discharge, peak-discharge differences, and total storm-event volume differences. Operationally, the forecasted streamflow corresponds to the scenario with the best historical performance, given the total amount of observed rainfall. The overall results suggest that the described approach is promising. However, there are still some cases in which the method leads to discharge underestimation. Considering the forecast uncertainty, the results show that it is possible to design flash floods alerts using this simple but robust methodology.