Property:CSDMS meeting abstract
From CSDMS
This is a property of type Text.
2
Much of the estimated 600 Mt of river sediment annually carried by the Ayeyarewady and Thanlwin River system (Myanmar) is delivered to the northern Andaman Sea. This area is influenced by strong tides, monsoon conditions, and periodic cyclones; however the processes that dominate dispersal of riverine material in the coastal ocean of this system have remained largely unquantified. The shelf exhibits a dramatic asymmetry of the surface morphology and sediment texture in the east – to – west direction, and recent field observations indicate that sediment accumulation rates increase toward the west. To explore the role that wave resuspension may play in these patterns, the SWAN (Simulating WAves Nearshore) model was implemented for the northern Andaman Sea, and run to represent both winter and summer time periods. The wave orbital velocities provided by SWAN were then analyzed to estimate the frequency of resuspension of fine-grained sediments throughout the study area. Results showed that wave-driven resuspension is much more frequent during the summer conditions which are characterized by the southwest monsoon; compared to during the northeast winds typical of the winter season. Additionally, the area fronting the Ayeryarwaddy Delta is subjected to energetic waves throughout both the summer and winter conditions, but wave energy decreases dramatically eastward toward the Thanlwin River mouth. +
A
NOAA’s National Geophysical Data Center (NGDC) develops and publicly distributes a wide variety of topographic and integrated bathymetric-topographic digital elevation models (DEMs), ranging from the global ETOPO1 and GLOBE, to high-resolution (~10-m cell size) coastal DEMs to support NOAA’s tsunami forecast and warning efforts. We have developed a prototype online tool, using an underlying THREDDs catalog, to view and extract the square-cell models in their native resolution and datums, subset by user extents, and output in netCDF, geotiff, xyz, or ESRI Arc ASCII formats. We have also implemented a command-line get request that bypasses the browser interface. Current models include the 1-minute ETOPO1, 30-second GLOBE topography, the 3 arc-second U.S. Coastal Relief Model (CRM) and Great Lakes Bathymetry, and the 24 arc-second Southern Alaska CRM. In the future, we will be expanding the catalog to include all of NGDC’s public DEMs, and are investigating ways to in-fill gaps between higher-resolution DEMs with data from coarser models. +
2
Natural channels are continuously changing their shape, where meanders and other complex configurations appear (e.g. bars, braided rivers, inner confluences, etc.). Channel evolution is strongly determined by the interactions occurring between its banks and the flow. These interactions also determine when a channel stabilizes, i.e. when its width remains constant. Current literature explains the stabilization of channels by the attainment of the equilibrium between sediment diffusion and gravity forces. However, the role of other potentially relevant processes is uncertain and needs to be addressed. Among them are secondary currents close to the banks and the spatial distribution of turbulence. Furthermore, the transition to steady-state banks is not fully understood. We explored these issues aiming to provide a better understanding of bank erosion and channel stability. To do this, we simulated a flatbed channel under 8 conditions, with Shields parameter spanning from 0.03 to 1.78. These simulations solved a 3D turbulent flow by carrying out Large-Eddy Simulations (LES) and the particles’ motion through a Discrete Element Method (DEM). We observed streamwise-aligned vortices appearing close to the banks, which were associated with high levels of TKE and shear stress, as well as flow spanwise velocity fluctuations. These fluctuations were mainly sweeps and ejections, which helped to dislodge sediments from the banks. Once detached, sediments could travel downstream. The role of the turbulence was also observed by separating the diffusive and advective components of the transport, where the initial bank erosion was dominated mainly by the former. Indeed, turbulence roughly explained 90% of sediment flux under erosion and bedload transport conditions. We conclude turbulent events increase shear stress close to the banks, promoting entrainment. Once the flow has transferred enough momentum to sediments, flow mean-velocity and fluctuations decrease. In this manner, shear stress decreases as the channel width increases. Eventually, shear stress reaches the threshold for transport close to the banks. Here, channel stabilization occurs. Notwithstanding that, stresses in the center of the channel are high enough to continue transporting sediments.
Nearshore hydrodynamic modeling necessitates extraordinary computational power to resolve the scales of motions relevant to coastal processes. Thus, coastal models make tradeoffs in the processes resolved. One common tradeoff is wave-averaging, whereby the evolution of bulk properties and statistics of wave fields are modeled. This contrasts the computationally more intensive wave-resolving models, whereby the time-varying motion of individual waves is directly output. However, complex nearshore dynamics are often driven by phenomena that cannot be directly derived from wave-averaged quantities, which limits the breadth of applications for wave-averaged models. Machine learning techniques provide a potential avenue to leverage the power of wave-resolving models for such applications at a lower computational cost. To this end, the wave-resolving, depth-integrated FUNWAVE-TVD modeling based on solving the Boussinesq equation is used in this study. The model was validated against the Dune3 dataset collected at Oregon State University corresponding to wave evolution and breaking in a cross-shore surf zone. A series of similar wave-resolving simulations using the FUNWAVE-TVD model were generated to create a training dataset corresponding to a one-dimensional planar beach under regular wave conditions. Two properties of interest, wave skewness and asymmetry, were calculated from the resulting wave-field and parameterized via the input wave conditions and bathymetry. Preliminary results show that even relatively simple ML models (neural networks and random forests) can provide drastic improvements to commonly used empirical models commonly employed by wave-averaging models, suggesting that ML-based parameterizations of nearshore wave properties provide a viable avenue for improving wave-averaged models. +
Neural networks (NNs) enable precise modeling of complicated geophysical phenomena but can be sensitive to small input changes. In this work, we present a new method for analyzing this instability in NNs. We focus our analysis on adversarial examples, test-time inputs with carefully crafted human-imperceptible perturbations that expose the worst-case instability in a model's predictions. Our stability analysis is based on a low-rank expansion of NNs on a fixed input, and we apply our analysis to a NN model for tsunami early warning which takes geodetic measurements as the input and forecasts tsunami waveforms. The result is an improved description of local stability that explains adversarial examples generated by a standard gradient-based algorithm, and allows the generation of other comparable examples. Our analysis can predict whether noise in the geodetic input will produce an unstable output, and identifies a potential approach to filtering the input that enable more robust forecasting. +
North Core Banks, a long (36-km), low (2.6-m mean elevation), narrow (~1200-m) barrier island in the Outer Banks of North Carolina, was inundated from the sound side and severely eroded by outwash during Hurricane Dorian (September 2019). As the fast-moving Category-1 hurricane moved offshore after a brief landfall at Cape Hatteras, winds shifted to the northwest, forcing a ~2.5-m surge onto the back side of the island. Deeply incised drainages were cut into the island as water ran from the washover platform to the ocean through gaps in the primary dune line, removing ~16% of the island volume. This style of storm impact is less common than typical ocean-side attack by waves and storm surge, and rarely modeled. Model simulations may provide insight into the fate of sands eroded during these unusual and difficult-to-measure events. We used the COAWST modeling system to simulate conditions during Dorian for a typical segment of the island using topography and landcover derived from pre-storm mapping using aerial imagery. The high-resolution (~2-m horizontal grid spacing) model was forced by output from a coarser-resolution model that provided water levels, incident waves, and alongshore currents. The model reproduced the steep cross-island water-level gradients inferred from high-water marks and wrack deposits, and generated washout channels ~2 m deep, cut through pre-existing low spots in the primary dune line. We evaluated model performance by comparing the simulated topography with post-storm topography derived from aerial imagery. The location, depth, and width of the simulated channels matched observations well, but the inland portion of the modeled channels were more linear than the observed dendritic drainages. Model simulations were sensitive to water-level forcing, sediment size, and vegetation patterns. The simulated channels extended into the surf zone and deposited sediments in relatively deep water. This transfer of sand from the island core to the nearshore has implications for barrier island evolution, and the ability to model it with COAWST demonstrates the generality of its morphology components.
A
Notice: Kim Picard is 1st author; Phil Hill 2nd author; Andrew Wickert 3rd author<br><br>This work aims to improve the late Quaternary stratigraphic framework for the outer shelf and slope of the Beaufort Sea and to assist in the assessment of geohazards, particularly those related to slope instability. Slope failures have been identified on the upper slope, but the age and triggers of slope failure are poorly understood. Existing conceptual models of late Quaternary stratigraphy of the Beaufort shelf and slope are quite generalized and based on a poorly constrained relative sea level curve. Sea level and stratigraphic modeling are used to test the relationships between glaciation, sea level and sedimentation. The results of the work suggest that glacio-isostatic effects cause the relative sea level (RSL) curve to vary significantly across the Beaufort Shelf particularly in the cross-shelf direction. Stratigraphic modeling with a variable RSL input successfully reproduces depositional patterns in the Mackenzie Trough including distinctive highstand and lowstand wedges and a retrogradational transgressive systems tract. However on the eastern shelf, more pronounced isostatic depression is required to match the known stratigraphy, suggesting deviation from the assumed ice loads or crustal properties in the model. Two outburst floods documented to have occurred in the region would have had a marked effect on shelf edge and slope sedimentation. Modeling suggests significant progradation of the shelf edge and rapid deposition on the slope and outer shelf at lowstand and in the early stage of transgression. +
2
Numerical models are effective and efficient tools for understanding the interactions among earth surface processes, including hydrological, biogeochemical, geomorphological, and ecological processes. These models with various complexities test hypotheses and make predictions within certain pre-defined model boundaries. These boundaries, on the one hand, reduce the complexity and noise of the system by ignoring the processes outside the boundary, which may only play a minor role in affecting the dynamics of the modeling system. On the other hand, some key processes that have a first-order control on the system dynamics may be unreasonably excluded from the modeling system. With the rapid growth of interdisciplinary researches, there is a more urgent need of revisiting the definition of the boundaries in the current numerical models to understand the gaps in bridging the boundaries between disciplines. This study is developed to meet this need. We investigated the models in the CSDMS model repositories by analyzing the process integration, boundary conditions, and spatial and temporal scales and summarized the potential gaps between boundaries that the current models present. This is the first study that conducts a comprehensive review of the models in the earth system modeling community, which provide insights for future model development and implementation across boundaries. +
Numerical models play a vital role in understanding river channel and floodplain evolution, yet their setup often requires extensive measured data. Maintaining continuity in monitoring fluvial geomorphology and sediment transport globally poses a significant challenge. This study introduced a remote sensing-based methodology for constructing and calibrating a reach scale 1D hydrodynamic numerical model, particularly suited for data-scarce regions. The effectiveness of this approach was assessed on the Elwha River in Washington.
The methodology employed a supervised image classification technique to extract a river mask, especially useful in areas with significant shadow pixels. Subsequently, channel attributes such as width, sinuosity, and slope were derived, and river segments with similar cross-sectional properties were identified using a multivariate change point approach, resulting in delineation of four distinct segments for the Elwha River. Next, hydraulic calibration of the numerical model accurately simulated water surface elevation (NSE: 0.93, PBIAS: -7%, RSR: 0.27). The sediment transport sub-model provided precise estimates of Suspended Sediment Concentration for mid-discharge values of 70 – 100 m3/s, associated with exceedance probabilities ranging from 0.4 to 0.04. Furthermore, the numerical model accurately reproduced channel deposition-erosion patterns estimated using publicly available aerial imagery from 2015 to 2017 (56 m vs. 48 m). These findings demonstrate the successful utilization of remote sensing datasets to supplement data requirements for numerical model setup and calibration, as well as to generate validation datasets. The methodology holds promise for accurately simulating hydromorphodynamic processes in both data-rich and data-scarce regions. +
Numerical simulation of fluvial morphodynamic processes can identify important dynamics at time and space scales difficult to observe in the field. However, simulations involving large spatial scales and/or the long timescales characteristic of morphodynamic processes are often untenable due to long simulation times. The morphological acceleration factor (morfac) applies a scalar multiplier to the sediment continuity equation, and is often applied in morphodynamic simulations to reduce computational time. While the use of morfac in coastal simulations is relatively common, its applicability in field-scale fluvial models is generally confined to steady-flow simulations over reach-scale spatial domains. Here we explore the viability of using morfac to simulate large-scale, long-term morphodynamics in a gravel-bed river. Using Delft3D to simulate a 60-day period with a significant discharge event in the Nooksack River, Washington, we systematically adjust morfac values (ranging from 5 to 20) to compare with a baseline condition of no acceleration. Model results suggest that morfac based modification of the inflow hydrograph time-series significantly alters downstream flood wave propagation. Higher morfac values result in greater flood-wave attenuation and lower celerity, reducing the morphological impact at locations further downstream. In general, relative error compared to the baseline increases farther downstream, due to this altered flood-wave propagation. Furthermore, even for the lowest morfac values absolute cumulative volume change errors are on the order of 10%, indicating that the use of morfac in fluvial simulations is best restricted to short-term and/or smaller-scale modeling efforts. Funded by the National Science Foundation. +
Numerous humanitarian and disaster relief missions require updated topography to provide time-critical support following events such as earthquakes, wildfires, tsunamis, or armed conflicts. Radar techniques are particularly advantageous over other methods (e.g., LiDAR) in these scenarios because they are insensitive to weather and lighting conditions, allowing data collection through clouds and smoke, or at night. Despite these advantages, current radar-based methods for generating topography face significant challenges related to data acquisition logistics, processing complexity, and the specialized expertise needed.
Inspired by recent advances in computer vision and monocular depth estimation, we present a novel approach to generate Digital Elevation Models (DEMs) from single Synthetic Aperture Radar (SAR) images using deep learning. Our method leverages a global dataset of open-source SAR-DEM image pairs to train multiple architectures, including Vision Transformers (ViTs) and fully convolutional networks. We evaluate various supervised and adversarial training strategies across a diverse range of Earth's landscapes. Our approach streamlines topographic reconstruction by working directly in ground coordinates and eliminating specialized pre-processing, making DEM generation more accessible. By utilizing open-source satellite radar data with a 6-day revisit time, our method enables topographic reconstruction at a significantly improved temporal resolution. +
Observations in coastal environments show that seabed resuspension can impact water quality and biogeochemical dynamics by vertically mixing sediment and water, and by redistributing material that has been entrained into the water column. Yet, ocean models that incorporate both sediment transport and biogeochemical processes are rare. The scientific community frequently utilizes hydrodynamic-sediment transport numerical models, but hydrodynamic-biogeochemical models ignore or simplify sediment processes, and have not directly accounted for the effect of resuspension on oxygen and nutrient dynamics.
This presentation focuses on development and implementation of HydroBioSed, a coupled hydrodynamic-sediment transport-biogeochemistry model that was developed within the open-source Regional Ocean Modeling System (ROMS) framework. HydroBioSed can account for processes including advection, resuspension, diffusion within the seabed and at the sediment-water interface, organic matter remineralization, and oxidation of reduced chemical species. Implementation of the coupled HydroBioSed model for different locations, including the Rhone River subaqueous delta and the northern Gulf of Mexico, have helped to quantify the effects of both sediment transport and biogeochemical processes. Results indicate that resuspension-induced exposure of anoxic, ammonium-rich portions of the seabed to the more oxic, ammonium-poor water column can significantly affect seabed-water column fluxes of dissolved oxygen and nitrogen. Also, entrainment of seabed organic matter into the water column may significantly draw down oxygen concentrations in some environments. Ongoing work focuses on how resuspension and redistribution of organic matter and sediment may influence oxygen dynamics in the Chesapeake Bay. +
Observations of the spatial and temporal evolution of thaw and soil moisture changes are needed to understand thermo-hydrologic dynamics in periglacial regions and to inform models that forecast changes in the Arctic. However, obtaining spatially and temporally distributed observations in the Arctic is difficult. Here we develop and investigate the use and accuracy of the parameter estimation algorithm in recovering soil physical properties. We tested our parameter estimation (PE) approach with synthetic data from a continuously modeled electric resistivity tomography transect and co-located synthetic temperature and soil moisture data. The results indicate that developed PE approach is able to identify synthetic porosities and thermal conductivities. +
Ocean waves are key drivers of erosion and cliff retreat along rocky coasts, doing so by delivering energy to the shore upon breaking. Wave energy attenuation increases with increasing distance from the location of breaking. As a result, breaking distance from the shore is one of the most important constraints on wave energy delivery to the coast. A primary factor influencing nearshore wave transformation and energy flux at the shore is shore morphology. We seek to evaluate local morphologic controls to better characterize wave energy delivery to the coast. Local wave climates are characterized utilizing NOAA datasets, and we incorporate the Coastal Relief Model to determine nearshore bathymetry and coastal morphology. We then perform shallow water wave transformations using linear wave theory to specify wave breaking locations along the shore. Here we present preliminary results that suggest that shore morphology, and specifically the gradient of the shore platform, is the dominant control on wave filtering and transformation along the West Coast of the United States. Ascertaining the role of shore morphology in controlling energy delivery to the shore is important for specifying the influence of shore steeping processes on wave transformation and energy delivery, as well as constraining and predicting coastal erosion and cliff retreat. +
A
Often a rivers discharge is calculated by constructing an empirical relationship between concurrent, direct measurements of river stage and discharge. In many remote parts of the world however technical and logistical challenges make building of such relationships difficult.
We test and present an alternative approach for use in remote Greenlandic Rivers. We used in-situ stage observations, but converted these measurements into estimates of discharge using a fluid mechanically based model (Kean and Smith, 2005; Kean et al., 2009; Kean and Smith, 2010). We first tested this approach against the one river in Greenland with a well-developed empirical stage- discharge relationship. Modeled relationships agreed well with the empirically derived relationship. We then used this same technique to aid in estimating discharge on two additional rivers in Greenland where only stage measurements were available. This technique presents an alternative option when other methods are logistically prohibitive. In the future this approach may also be useful to aid in estimating river discharge from space. +
2
On a broad scale climate controls the distribution of biomes and sets an upper limit for woody plant canopy cover. During last glacial cycle that peaked ~18,000 years (B.P.) in the Late Pleistocene, the southwestern United States was much wetter and cooler than in the Holocene (last 11,000 years) and today. Since the Last Glacial Maximum, wetter and cooler climate in most arid semiarid regions has generally transitioned to drier and warmer conditions, establishing their characteristic (i.e., today’s native) ecosystems and fire regimes 3,000 - 5,000 years B.P. We use the Landlab earth surface modeling toolkit to explore the implications of the climate since the late Pleistocene on ecosystem patterns, using a calibrated model for conditions prior to the Euro-America settlers. Climate is constructed based on paleoclimatic proxies and weather station data. The controls of seedling dispersal strategies of plants and water availability as mediated by aspect are discussed. +
On densely populated deltas, the tendency for river channels to catastrophically avulse poses a hazard to human life and property. Previous work has shown that river avulsions preferentially occur around a spatial node with a distance from the shoreline that is controlled by backwater hydrodynamics, the interplay of dynamic river discharge and standing water near the shoreline. Our ability to forecast the location of future avulsion hazards is limited, however, because avulsions are relatively rare and many deltas are experiencing drastic changes in river discharge and sea level due to land-use and climate change. Building upon previous work, we present a predictive model of delta-lobe morphodynamics and repeated avulsion that is applicable to deltas over a range of spatial scales, sediment supplies, flood regimes, and relative-sea-level-rise conditions. In our model, delta lobes build on top of one another, demonstrating a distribution of avulsion lengths that is sensitive to flow regime and relative sea-level change. Variable flood regimes lead to a consistent avulsion length when low flows (less than bankfull) and high flows (greater than bankfull) compete to intermittently fill and scour portions of the backwater reach. The avulsion node is a spatial maximum in channel superelevation set by the downstream extent of low-flow deposition between erosive high-flows, and in general channels avulse farther upstream when high-flow events are more extreme and more frequent. Relative sea-level rise leads to a more variable avulsion node, driven by intermittent retreat and advance of the delta shoreline as the river periodically shifts the distribution of sediment. If rise rates are sufficiently high to sequester all sediment upstream of the river mouth, avulsions occur progressively farther upstream or not at all. These results have implications for the forecasting of avulsion hazards on modern deltas undergoing relative sea-level rise and human management.
A
On earth, landscape morphology is mainly controlled by rivers evolutions and their interactions with hillslopes. But hydrographic network may be re-organized by stream capture and modify deeply the relief. This transition may be induced by several mechanisms (diversion, headward erosion, avulsion, or subterranean filling up). It has interested numerous scientists since a long time (Davis 1895, Blache 1943, Lesson-Quinif 2001 & Le Roux-Harmand 1997-2009…). Here we focus on stream piracies by headward erosion, when an actively eroding low level stream (called the captor) encroaches on the drainage of a nearby stream flowing at a higher level (called the diverter) and diverts part of the water of the higher stream.
During the last decades, several landscapes evolution models (LEM) have been developed to quantify the topography evolution with diffusion and advection equations. These models play an important role in sharpening our thinking to better understand the interaction between landscape evolution processes. LEM were developed basically to simulate erosion, tectonic and climate at different scales of time and space. But, these models were not designed to describe specific mechanisms as the stream capture. It’s one of the aims of this work to evaluate LEM for this purpose.
In this paper, we develop a 1D model based on LEM equations to investigate the stream piracy by headward erosion responses to climatic or tectonic changes. This model incorporates the most common equations used in quantitative geomorphology; diffusion in hillslope, advection in river (detachment-limited mode) and an inequality based on slope and drainage area for the limit between these two domains (Montgomery and Dietrich, 1988). First, simulations on analytical cases highlight the stream head progression mechanism, and the results indicate that this progression rate is mainly controlled by the slope at the captor source. Consequently, the aggradation of the diverter or (and) the incision of the captor accelerate the process. Then, a predictive study with an improved version of GOLEM (software developed by Tucker & Slingerland in 1994) on the Meuse basin shows that several piracies may probably occur in the future. A comparison with the 1D model gives similar results.
The simplicity and the flexibility of the 1D model allow complex simulations in the Meuse basin taking into account: lithological differences of outcropping layers, Meuse deposition tendency, etc. Once the 2D simulations or topography analysis locate potential captures, 1D simulation may intensively be used, as it presents many advantages; weak execution time, simple limits conditions setting, less time for data preparation, etc. Consequently, a sensitivity analysis to estimate piracies ages is realized with the developed 1D model.
2
On low-lying coastlines, sand (and/or gravel) washed landward of a beach during storms creates barrier landscapes. This ‘overwashed’ sand also tends to maintain barrier elevation in the face of rising sea level, and barrier width in the face of an eroding shoreline. However, from the point of view of coastal communities, the storm processes that deposit the sand, as well as the sand itself, presents hazards that need to be mitigated, or even disasters that need to be recovered from. The strategies typically chosen to mitigate storm hazards and recover from storm impacts typically involve attempts to prevent overwash processes (e.g. by building and maintaining large dunes) or to undo the effects of overwash processes (e.g. bulldozing overwashed sand off roads and using it to rebuild dunes). Although these mitigation and recovery strategies seem necessary in the short term, they can reduce coastal resiliency in the long term, by tending to make a barrier lower (relative to rising sea level) and narrower (in the case of an eroding shoreline). A lower and narrower barrier is more vulnerable to sunny day flooding and severe storm impacts.
The newly developed CoAStal Community-lAnDscape Evolution (CASCADE) model couples physical processes (storm erosion and sediment redistribution, dune growth, sea-level rise, shoreface and shoreline change, and gradients in alongshore sediment flux) and the effects of management strategies (e.g. overwash removal and dune maintenance, highway relocation, and beach nourishment). Using this model, we examine the outcomes, over decades, of the coupling between natural dynamics and commonly employed management strategies. Modeled outcomes depend on sea-level-rise rate, storm sequences, and initial barrier topography, and they range from developed barrier systems that can be sustained for over a century before becoming uninhabitable (effectively drowned), to scenarios in which highways and/or communities need to be abandoned within decades. Subsequent barrier recovery depends on the final state of the developed system before abandonment, as well as stochasticity in the timing of storms. When different management strategies are employed at different locations alongshore, their effects are coupled via the redistribution of sediment along a curved coastline.
We are also using CASCADE in a participatory modeling collaboration involving managers and planners with the Cape Hatteras National Seashore and the North Carolina Department of Transportation, as well as community representatives. Together, we will examine the range of outcomes, under different climate scenarios, of the strategies being considered for managing a critically threatened transportation corridor along a barrier within the National Seashore.
One possible human response to climate change and other environmental stresses is migration. However, migration is complex, multi-causal phenomenon, and the complexity of human migration poses a challenge for researchers who aim to study the effects of environmental changes on population mobility. This project aims to understand how changing environmental conditions and livelihood opportunities impact migration decisions in coastal Bangladesh. An original agent-based model (ABM) that combines stylized environmental change dynamics with livelihood is developed to understand how these dynamics impact migration decisions as well as what feedbacks may exist between them. The ABM is constructed such that agents represent households, consisting of individuals, within a single origin community. At each step of the model, an agent will first assess the expected utility of its different options within the community, including doing nothing, seeking employment internal to the community, and investing in non-agricultural livelihood options. After assessing livelihood options internal to the community, households with sufficient wealth and a sufficient number of family members will decide whether or not to send a household member as a migrant, also based on expected utility of a migration trip. The model’s representation of natural processes will be simulated in the form of drought, modeled stochastically, that impacts crop yields and crop-associated income. In this initial version of the ABM, agent decision-making is based on simple utility maximization. Future work will incorporate more complex decision-making theories into the model, as well as different destination locations and the possibility of return migration. +
