Back to coastal-lakes atlas

Smithsonian fellowship · Scientific framework

From inherited lagoons to engineered waterscapes

Human–environment interactions along Egypt’s Mediterranean coast through the MEDIBA archive

Hader Sheisha · Proposed 24-month residency · Updated 5 October 2026

01 · Questions and scales

Research problem and objectives

Egypt’s Mediterranean coastal lakes record interactions among Nile discharge, sediment supply, marine exchange and relative sea-level change. Their inherited configurations shaped settlement, resource use and navigation, while successive interventions modified lagoon hydrology (Stanley et al., 1996; Flaux et al., 2012; Giaime et al., 2022). Roman canals connected Mareotis basins and settlements; later irrigation, river regulation and drainage expanded the scale of modification (Crépy and Boussac, 2021; Flower et al., 2001, 2009). Yet transformation is better documented than its rates, temporal sequence and causal pathways.

During a 24-month Smithsonian residency, I will ask how inherited Holocene lagoon configurations conditioned later human–environment interactions, and whether recent ecological change became faster, more persistent or compositionally different. Burullus and Manzala are the principal systems; Mariut provides a comparison where preservation and chronology permit (Figure 2).

01

Environmental inheritance

Reconstruct lagoon configurations and evaluate how inherited connectivity conditioned subsequent responses.

02

Historical interactions

Establish the sequence of interventions, environmental change and social reorganisation using independent evidence.

03

Ecological transformation

Test whether recent turnover exceeded earlier variability at comparable temporal resolution.

Figure 1. Research roadmap. Four operational aims connect archive qualification, inherited lagoon configurations, historical transitions and tests of recent ecological change. The temporal windows organise the research; increasing human influence, acceleration, synchrony and thresholds remain hypotheses to evaluate. Meeting and seminar frequencies represent the proposed collaboration rhythm.

Select the figure to enlarge.

The working hypothesis is that inherited connectivity and local interventions produced divergent responses to shared regional pressures. Competing explanations include episodic reorganisation and apparent acceleration caused by sampling or chronological differences. Acceleration, synchrony and threshold responses remain hypotheses, not starting assumptions.

Table 1. Nested temporal windows and their analytical purposes
Temporal windowResearch focusEvidence and interpretation
Mid-to-late HoloceneInherited environmental configurationsFreshwater, marsh, restricted-lagoon, brackish and marine-influenced states; hydrogeomorphic boundary conditions.
Last ~2,000 yearsReciprocal human–environment interactionsRoman, Late Antique, medieval/Islamic, Ottoman and modern settlement, navigation, resource use and water management.
Last ~500 yearsHistorical transition and hydraulic managementChanging canals, irrigation, drainage and reclamation; Ottoman and nineteenth-century evidence where independently dated.
Last ~150–200 yearsRecent ecological transformationRate, direction, persistence, synchrony and lags; fine-scale inference only where chronology and sampling support it.

These windows support different questions and achievable resolutions. Historical conditions provide a changing reference envelope rather than an undisturbed baseline. The 1964–1965 Nile-regulation transition is an explanatory window to evaluate, not an imposed breakpoint or independent age control.

02 · Spatial framework and collections

Archive recovery and qualification

Figure 2. MEDIBA core locations and coastal-lagoon context. Latest map supplied in the research proposal. Burullus and Manzala are the principal systems; Mariut is comparative. Modern imagery and generalised elevation provide spatial context, not historical shorelines. Coordinates, datum assumptions and imagery sources are retained within the figure.

Select the figure to enlarge.

The MEDIBA catalogue documents 87 northern Delta borings (Stanley, McRea and Waldron, 1996). Reconnecting their geological framework with short-core, surface-sample and historical evidence will extend the archive’s value for investigating lagoon transformation.

Inventory will reconnect long cores (Stanley, McRea and Waldron, 1996) with approximately 100 reported short cores and 200 surface samples documented outside the main monograph (Randazzo, 1992; Loizeau and Stanley, 1993, 1994; Bernasconi and Stanley, 1994; Siegel et al., 1994). I will reconcile identifiers, depths and sample histories across logs, analytical sheets and X-radiographs, then screen up to 24 levels from candidate Burullus and Manzala sequences. During months 3–6, continuity, preservation and chronological potential will determine selection (Figures 3–4).

Existing measurements will anchor reconstruction. I will recover pollen, diatom, macrofossil and aquatic-microfossil matrices, retaining taxonomy, counts and preservation metadata, and use targeted microscopy to assess surviving material. Lithology, grain size and geochemistry will establish depositional context. CASSARINA and MELMARINA comparisons require verified provenance and explicit cross-core uncertainty (Birks et al., 2001; Flower et al., 2009).

At the National Museum of Natural History and Museum Support Center, identifiers, coordinates, depths and sample histories will be reconciled across notebooks, logs, analytical sheets, photographs and available X-radiographs. Retained sand fractions, grain mounts, clay preparations, shells and other processed materials will be assessed for the analyses they can actually support. Their existence does not establish the survival of untreated sediment or complete proxy assemblages.

Figure 3. Preliminary screening of the MEDIBA coastal-lagoon archive. Archived upper sections are shown alongside depths to the shallowest published radiocarbon controls and sector maps. Shaded intervals above dated horizons require chronological and preservation screening; their thickness establishes neither age nor continuous accumulation. Ages are reported as conventional, uncalibrated radiocarbon years BP. Historical lagoon boundaries and engineering overlays are schematic and are not surveyed historical shorelines.

Select the figure to enlarge.

Figure 4. Core qualification and analytical strategy. Preliminary ratings distinguish chronology, apparent physical continuity and proposed analytical role. A = good, B = moderate or qualified, C = poor or problematic for the stated objective. Roles remain conditional on source verification and proxy preservation. Recovery or storage breaks are not automatically depositional hiatuses, and reworking interpretations require independent support. The provisional S78 comparison is restricted to approximately 4.5–6.5 m.

Select the figure to enlarge.

Table 2. Provisional ratings and core roles corresponding to Figure 4
LagoonCoreChronology / continuityProvisional roleQualification required
ManzalaS5A / APrimary long-term candidateVerify recovery, age ordering and proxy preservation.
ManzalaS2A / BReplication / sensitivity candidateEstablish chronological overlap with S5 and comparable proxy coverage.
BurullusS53A / AHolocene ecological reference candidateReconcile later AMS chronology and pollen evidence with legacy controls.
BurullusS52B / CPossible late-Holocene / historical bridgeSeparate fragmented recovery from demonstrated depositional hiatuses.
BurullusS48B / CProcess / taphonomy comparisonTest sedimentary discontinuities, reworking and proxy preservation.
Mariut sectorS78B / BInterval-specific Holocene comparisonPrioritise approximately 4.5–6.5 m; verify integrity and age controls.
Mariut sectorS79C / CFacies / reworking comparisonAssess shell taphonomy, age ordering and recovery before ecological interpretation.
Mariut sectorS81B / CHistorical-top pilotAssess upper-section survival; do not assume a continuous Holocene sequence.

Selection combines chronological suitability, physical continuity and proxy preservation. S78 is considered only as an interval-specific comparison, provisionally around 4.5–6.5 m. If historical intervals cannot be qualified, independently dated comparative records will replace unsupported continuity claims. A documented negative screening outcome remains part of the archive assessment.

03 · Evidence to analysis

Figure 5. Provisional study design: potential analyses and their contributions. The framework links chronological control, biological assemblages, sediment characterisation and selected shell analyses to environmental reconstruction and historical change. Analytical selection follows archive qualification, preservation and permissions. Individual proxies constrain interpretations jointly; they do not uniquely determine salinity, connectivity or hydrological forcing. The accompanying table specifies eligibility and the independent historical-evidence component.

Select the figure to enlarge.

Provisional fellowship analyses framework

Existing measurements provide the analytical foundation. New observations address specific gaps revealed by the archive audit and a pilot of up to 24 levels from candidate Burullus and Manzala sequences. The framework below distinguishes routine synthesis from analyses contingent on retained material, preservation, permissions and facilities; it does not prescribe a complete new proxy suite for every core.

Table 3. Potential analyses, conditions and benefits
AnalysisPlace in the designEvidence requiredScientific contribution
Archive and legacy-data recoveryCore researchLogs, analytical sheets, X-radiographs, sample histories and original matricesRe-establish provenance, observation units and usable intervals; avoid counting publications as independent datasets.
AMS dating and Bayesian age modelsTargeted after qualificationSuitable dated material and audited legacy controlsConstrain selected transitions, identify age inconsistencies and propagate chronological uncertainty.
Pollen and non-pollen palynomorphsRecovery plus targeted microscopyCountable, adequately preserved assemblagesReconstruct vegetation, marsh dynamics, freshwater vegetation and indicators of catchment disturbance.
Diatom assemblagesPreservation-dependentTaxonomically resolved assemblages and ecological informationEvaluate freshwater–brackish–marine shifts alongside other evidence for hydrological connectivity.
Ostracods, foraminifera and molluscsMaterial-dependentTaxonomy, abundance, preservation and depositional contextAssess aquatic habitats and marine exchange; distinguish ecological signals from transport and reworking.
Grain size, mineralogy and XRF / geochemistryQuestion-led and material-dependentComparable measurements and suitable surviving sedimentCharacterise depositional conditions, sediment supply and possible provenance changes.
Existing grain mounts and XRD materialArchive-dependentIdentifiable retained preparations and analytical historiesReassess sediment components and transport pathways; processed fractions cannot replace untreated sediment for every proxy.
Selected shell isotopes and elemental chemistryOptional, subject to suitabilityTaxon, preservation, context and analytical accessTest specific hydrological interpretations; avoid assuming shell chemistry uniquely identifies a water source.
Historical and archaeological evidenceCore researchIndependently dated and georeferenced sourcesEstablish intervention, settlement and resource-use sequences independently of sediment chronology.
Rate of change and trajectory comparisonQualified records onlyComparable effective resolution, assemblages and chronology ensemblesTest acceleration, persistence, direction and cross-lagoon timing with explicit sensitivity analyses.
QualifyMaterial, provenance, preservation
ConstrainChronology and effective resolution
ReconstructEnvironmental states and historical sequences
CompareTrajectories, alternatives and uncertainty

Each record retains its taxonomy, original counts where available, units, sample depth, source and preservation metadata. Cross-core correlation is treated as an inference with uncertainty. Spatially distinct cores are not merged into a continuous sequence solely because their proxies resemble one another.

04 · Temporal control

Chronology and effective resolution

Poorly dated upper sections are a specific target (Figure 3). Sediment above the shallowest radiocarbon control is not necessarily old or continuously deposited. S53 demonstrates the potential: Bernhardt, Horton and Stanley (2012) analyzed 74 pollen samples from its upper 10.5 m, using seven AMS dates and a Casuarina marker, compared with two dates in the original catalogue. The fellowship will test where comparable chronological development is feasible.

I will audit radiocarbon determinations for material, depth, laboratory information and reworking, distinguishing conventional from calibrated ages. Targeted AMS dating will prioritize terrestrial remains and boundaries critical to interpretation; shell ages require assessment of reservoir effects. Recent chronologies will use recovered lead-210 activities, supported activity, density and measurement errors, with available caesium-137 and americium-241 evidence. Weak Egyptian fallout signals make independent controls particularly important (Appleby et al., 2001). Model comparison will assess supply, accumulation and mixing assumptions; Bayesian age–depth ensembles will propagate uncertainty where inputs permit (Aquino-López et al., 2018). Hiatuses will remain explicit.

Radiocarbon and historical intervals

Audit dated material, depth, laboratory information, age ordering and reworking. Prioritise suitable terrestrial remains and interpret shell ages only after considering reservoir effects. Preserve hiatuses explicitly.

Recent sediment chronologies

Recover original ²¹⁰Pb activities, supported activity, density and measurement errors. Assess available ¹³⁷Cs, ²⁴¹Am and independent markers; compare model assumptions before accepting recent age estimates.

Sample spacing, accumulation, mixing, preservation and dating uncertainty determine effective temporal resolution. Centimetre-scale slicing alone does not establish decadal resolution. Legacy records end at their sampling dates; they cannot document subsequent restoration without newer evidence. Age–depth ensembles will carry chronological uncertainty into rate estimates, synchrony and lag comparisons where the inputs permit.

05 · Independent historical evidence

Human–environment interactions

A georeferenced register will integrate archaeological surveys, historical maps, documentary studies and accessible primary records. Roman and medieval settlement and harbour evidence will be compared with Ottoman and nineteenth-century accounts of canals, irrigation and land use, followed by engineering, drainage and fisheries records. Entries will retain dates, locations and sources; evaluation will consider mapping uncertainty, survey purpose and reporting changes.

Historical evidence will be assembled independently of sediment chronology. Interventions under investigation will not also serve as age anchors. Inferences about environmental constraints on societies require independent occupation or resource-use evidence; documentary silence does not establish absence of activity.

Table 4. Historical-source register and inference controls
FieldInformation retained
Source and provenanceArchive or publication, author/producer, identifier and evidence type
Temporal supportEvent or occupation range, dating basis, uncertainty and possible reporting lag
Spatial supportOriginal location, georeferencing method and uncertainty; historical shoreline context
Process representedSettlement, navigation, canals, irrigation, drainage, reclamation, fisheries or resource use
Interpretive limitsSurvey purpose, coverage, preservation, reporting change and competing explanations

Roman and medieval evidence will be compared with Ottoman and nineteenth-century material, including hydraulic interventions associated with the Muhammad Ali period where source dating and spatial coverage permit. Historical period labels organise the evidence; they do not impose ecological boundaries. Environmental constraints on societies require independent settlement or resource-use evidence, while human forcing requires documented interventions consistent with the observed mechanism.

06 · Quantitative tests and interpretation

Testing ecological change and divergent trajectories

Lake-specific reconstructions will distinguish hydrological and sedimentary changes from ecological turnover. Competing explanations will be evaluated through temporal ordering, spatial correspondence, independent proxies and plausible mechanisms. Human forcing requires interventions consistent with the observed response; environmental influence on societies requires evidence linking altered conditions to settlement or resource use. Correlation alone will not establish causation.

Assemblage rate-of-change analysis will use temporal standardization and sensitivity tests for counts, dissimilarity and bin width (Mottl et al., 2021). GAM derivatives will assess trends in adequately resolved proxies; change-point models will identify sustained shifts; multivariate trajectories will distinguish direction from speed. Chronological ensembles will carry age uncertainty into rate estimates, synchrony and lag comparisons. Recent intervals will be coarsened to match older records. Sample spacing, mixing and dating uncertainty will determine effective resolution; sparse records will support phase comparisons rather than fine-scale rates. Persistence and recovery will be evaluated separately from turnover magnitude.

Table 5. Quantitative comparisons and their evidential requirements
MethodEligibility and controlsInterpretive purpose
Assemblage rate of changeTemporal standardisation; sensitivity to counts, dissimilarity and bin widthTests turnover while controlling for unequal sample support.
GAM derivativesAdequately resolved proxy series; uncertainty in fitted trendsDistinguishes periods of increasing or decreasing change from uncertain trends.
Change-point analysisSufficient observations and sensitivity to model settingsIdentifies sustained shifts without imposing historical intervention dates.
Multivariate trajectoriesComparable assemblages and consistent taxonomic treatmentSeparates the direction of ecosystem change from its speed.
Synchrony and lag comparisonAge ensembles and sufficient temporal resolutionEvaluates shared timing or delayed responses without overstating chronological precision.
Persistence and recoveryExplicit reference intervals and observation coverageAssesses duration and return trajectories separately from turnover magnitude.

Recent series will be coarsened to the temporal support of older records, with sensitivity to bin width, counts, taxonomic harmonisation, dissimilarity and chronology. Sparse records will support phase comparisons rather than artificially precise rates. A turnover peak alone does not establish a threshold, loss of resilience, an alternative stable state or hysteresis.

For palaeoecology and geoarchaeology, the project will evaluate how environmental inheritance and successive interventions produced divergent lagoon trajectories. For environmental management, it will identify documented reference conditions and distinguish historical variability from persistent transformation. My research on Nile landscapes and socioecological resilience, together with my palaeoecological work at the University of Bergen, provides the foundation. The fellowship will extend this work into historical coastal systems and establish reusable collections-based evidence for my long-term research programme in Egypt.

07 · Work programme and outputs

Figure 6. Proposed 24-month fellowship workflow. Archive screening and staged synthesis connect Holocene reconstruction, historical transitions and resolution-standardised tests of modern change. Outputs are targeted at month 7, months 16–18, months 18–20, month 21 and months 22–24. Publication labels denote planned manuscript development or submission, not guaranteed publication dates. Quantitative outcomes depend on qualified records and adequate temporal support.

Select the figure to enlarge.

Twenty-four-month research programme

Months are counted from the fellowship start date. Work packages overlap, and synthesis proceeds throughout the programme.

Table 6. Work packages and milestones corresponding to Figure 6
MonthsWork packageMilestone or output
1–6Archive reconstruction, chronological screening and pilot re-analysisQualify candidate intervals during months 3–6 using chronology, continuity and proxy preservation.
7First consolidation and archive releaseArchive and methodological report; first manuscript component and documented archive consolidation.
8–16Targeted Holocene re-analysis and environmental-state reconstructionAudited proxies and age models; reconstruction of inherited lagoon states.
10–18Historical-transition reconstructionIntegrate independently dated environmental, archaeological, documentary and cartographic evidence.
16–18Holocene review and gap fillingHolocene environmental-state synthesis and manuscript preparation.
18–20Historical-transition review and completionHistorical human–environment paper or report.
20–24Modern transformation and resolution-standardised change analysisModern-change and methodological manuscript targeted at month 21 where support permits; sensitivity analysis and revision continue through month 24.
22–24Final cross-lagoon synthesis and completionIntegrated interpretation, remaining manuscripts, archive inventory, reusable data and code, and atlas updates.
Month 7

Archive and methodological report; first manuscript component and archive consolidation.

Months 16–18

Holocene environmental-state synthesis and publication preparation.

Months 18–20

Historical human–environment paper or report reviewed and completed.

Month 21

Modern ecological-change and methodological manuscript, where temporal support permits.

Months 22–24

Cross-lagoon synthesis, remaining manuscripts, inventory, documentation and atlas updates.

Decision points during months 3–6 qualify candidate records; months 12–14 evaluate age models and analytical coverage; months 18–21 integrate reconstructions and comparisons. Website updates follow the month-7, months-16–18, months-18–20 and months-22–24 synthesis points. Data are presented here as on-page tables, without document or spreadsheet attachments or export controls.

The core design requires no new Egyptian field campaign. If only one lake supports quantitative comparison, the other will provide contextual evidence. If temporal support is insufficient for acceleration or lag tests, the project will report phase-level change and the limits of inference.

08 · Collections-based research

Smithsonian residence and collaboration

Dr Brian T. Huber, Research Geologist and Curator of Foraminifera in the Department of Paleobiology, National Museum of Natural History, is the proposed principal adviser and host. He has expressed willingness to advise and facilitate archive access if the fellowship is awarded. Jonathan Wingerath is the collections contact. Specific holdings, analytical suitability and permissions will be resolved with the custodians.

Residence is essential for repeated examination of MEDIBA materials alongside original documentation at the National Museum of Natural History and Museum Support Center. Working with collections staff will resolve sample histories and discrepancies unavailable from publications. The project will return verified identifiers, analytical histories and a reusable digital inventory to the Smithsonian.

Research will use authorised collections and documentation, Smithsonian Libraries and Archives resources, and approved microscopy, imaging and computing facilities. Destructive sampling, loans, reproduction, restricted documents and unpublished external datasets require permission from their custodians. CASSARINA and MELMARINA comparisons require verified provenance and access through the responsible repositories or collaborators.

09 · Sources

Selected methodological and regional references

  1. Appleby, P. G., et al. (2001). Radiometrically determined dates and sedimentation rates for recent sediments in nine North African wetland lakes (the CASSARINA Project). Aquatic Ecology, 35, 347–367. https://doi.org/10.1023/A:1011938522939
  2. Aquino-López, M. A., Blaauw, M., Christen, J. A., & Sanderson, N. K. (2018). Bayesian analysis of ²¹⁰Pb dating. Journal of Agricultural, Biological and Environmental Statistics, 23, 317–333. https://doi.org/10.1007/s13253-018-0328-7
  3. Birks, H. H., et al. (2001). Palaeolimnological responses of nine North African lakes in the CASSARINA Project to recent environmental changes and human impact detected by plant macrofossil, pollen, and faunal analyses. Aquatic Ecology, 35, 405–430. https://doi.org/10.1023/A:1013265310922
  4. Flaux, C., El-Assal, M., Marriner, N., Morhange, C., Rouchy, J. M., Soulié-Märsche, I., & Torab, M. (2012). Environmental changes in the Maryut lagoon (northwestern Nile delta) during the last ∼2000 years. Journal of Archaeological Science, 39, 3493–3504. https://doi.org/10.1016/j.jas.2012.06.010
  5. Flower, R. J., et al. (2001). Recent environmental change in North African wetland lakes: diatom and other stratigraphic evidence from nine sites in the CASSARINA Project. Aquatic Ecology, 35, 369–388. https://doi.org/10.1023/A:1011984627760
  6. Flower, R. J., et al. (2009). Sediment distribution and accumulation in lagoons of the Southern Mediterranean Region (the MELMARINA Project) with special reference to environmental change and aquatic ecosystems. Hydrobiologia, 622, 85–112. https://doi.org/10.1007/s10750-008-9677-5
  7. Giaime, M., et al. (2022). Holocene evolution and signature of environmental change of the Burullus lagoon (Nile Delta) deciphered from a long sediment record. Palaeogeography, Palaeoclimatology, Palaeoecology, 590, 110861. https://doi.org/10.1016/j.palaeo.2022.110861
  8. Mottl, O., et al. (2021). Rate-of-change analysis in paleoecology revisited: A new approach. Review of Palaeobotany and Palynology, 293, 104483. https://doi.org/10.1016/j.revpalbo.2021.104483
  9. Stanley, D. J., McRea, J. E., Jr., & Waldron, J. C. (1996). Nile Delta drill core and sample database for 1985–1994: Mediterranean Basin (MEDIBA) Program. Smithsonian Contributions to the Marine Sciences, 37, 1–428. https://doi.org/10.5479/si.01960768.37.1

Additional archive and historical sources cited in the proposal include Randazzo (1992), Loizeau and Stanley (1993, 1994), Bernasconi and Stanley (1994), Siegel et al. (1994), Bernhardt, Horton and Stanley (2012), and Crépy and Boussac (2021). Full provenance accompanies the relevant archive records.