# Egyptian coastal lakes and human societies
## Review protocol and synthesis architecture

**Version 1.1 · Dated amendment and alignment: 28 September 2026**

This protocol aligns the review with the fellowship scientific core: a collections-led investigation of how dynamic Holocene lagoons became historically managed and modern engineered socio-ecological systems. It replaces the earlier broad coastal-review architecture for the active project. The original frozen statement remains archived; the changes below are prospective research rules, not evidence that additional searches or analyses have been completed.

## 1. Review question

When, through which pathways and at what defensible temporal resolution did Egypt’s Mediterranean coastal-lagoon system take its modern form? How did changing waterscapes and human societies influence one another through the Holocene, particularly during Roman, Late Antique, medieval/Islamic, Ottoman and modern periods? Does recent ecological turnover exceed earlier variability after chronological uncertainty, mixing, preservation and unequal sampling are considered?

## 2. Geographic and institutional scope

Burullus and Manzala are the central analytical cases. Mariut provides a historical and geomorphic comparison and enters quantitative analyses only if accessible material, suitable proxies and defensible chronology are established. Edku extends the regional literature comparison; Bardawil is a contrasting lagoon, never an assumed unaffected control. Neither adds a compulsory sampling programme. Red Sea ecosystems and nationwide coastal forecasting are outside the fellowship’s core analysis.

The Smithsonian MEDIBA legacy associated with Daniel Jean Stanley supplies the proposed collections basis. Adviser agreement, present-day custody, access, digitisation permissions and sampling restrictions require confirmation. CASSARINA and MELMARINA records are complementary archives, not presumed Smithsonian holdings. Published catalogues establish historical work, not current access. The base plan requires no new Egyptian coring campaign.

## 3. Nested temporal framework

| Window | Purpose | Interpretation rule |
|---|---|---|
| Last approximately 150–200 years, with historical evidence from ca. 1800 CE | Principal test of ecological rates, mechanisms and lags | Sub-decadal to decadal inference only where effective resolution supports it |
| Last approximately 500 years | Bridge between inherited configurations and expanding hydraulic intervention | Neither pristine nor stationary reference conditions |
| Last approximately 2,000 years | Reciprocal waterscape–society interactions through Roman and later periods | Period labels organize evidence; they do not impose environmental boundaries |
| Mid-to-late Holocene | Lagoon formation, distributary history and earlier human–environment context | Coarse long records are not pooled directly with recent records for rate comparisons |

Retain 27 September 2026 as the original review search cutoff unless a dated search extension is recorded. The 28 September amendment changes scope and interpretation, not the date on which literature searching is claimed complete. Core termination dates constrain each reconstruction independently; older cores cannot evaluate subsequent rehabilitation.

The 1964–1965 hydrological transition is an explanatory hypothesis window, not a compulsory change point or chronological anchor. Other engineering, political and management dates will be treated in the same way. Intervals around interventions may guide comparison only after independent chronology has been established.

## 4. Questions and competing explanations

1. When did modern hydrological configurations and ecological assemblages emerge, and were their transitions synchronous within or between lakes?
2. Does recent turnover exceed earlier variability under comparable temporal support, or does apparent acceleration reflect denser sampling and better dating?
3. How did freshwater availability, salinity, navigability, marsh expansion and sedimentation affect settlement, agriculture, fisheries and transport?
4. How did canal management, drainage, reclamation, cultivation and resource use modify hydrological connections and ecological conditions?
5. Do differences among sectors reflect regional forcing, local geomorphology, management history or their interaction?
6. Is there defensible evidence of ecological lags, persistence, reorganization or recovery?

Test both directions of interaction without presuming an inevitable progression from adaptation to environmental control. Environmental constraints do not determine social outcomes; institutions and practices may mediate them. Chronological coincidence does not establish a mechanism. Statistical shifts alone cannot demonstrate a causal threshold, alternative stable state or hysteresis.

## 5. Evidence streams and search plan

Maintain linked sedimentary, ecological, archaeological, documentary, cartographic and modern observational streams. Search the original environmental databases and institutional repositories, with backward and forward citation tracking. Add targeted archaeological catalogues, historical scholarship and identifiable map or documentary collections for the northern Delta. Searches and translations in English, Arabic and French will retain original terms and provenance where feasible.

Combine lake-name variants (Burullus/Borollos, Manzala/Manzalah, Mariut/Maryut/Mareotis) with process terms and temporal terms. Environmental terms include lagoon, palaeochannel, sediment, pollen, diatom, salinity, chronology, drainage and reclamation. Historical terms include Roman, Late Antique, medieval, Islamic, Mamluk, Ottoman, settlement, harbour, canal, irrigation, fishing and land use. Ancient or variant place names must be geographically verified before assignment.

For every search record database or repository, date, exact query, filters, results and export identifier. The historical search is a required extension, not a completed systematic corpus. Use transparent environmental evidence-map reporting, documented screening decisions and explicit limitations where resources prevent independent duplicate screening. Do not claim registration or systematic-review completion without a recorded action.

## 6. Eligibility and appraisal

Include sources that provide traceable evidence for a relevant place, period and variable, or independently constrain a proposed mechanism. Retain spatial studies for spatial mechanisms; do not treat them as time series. Retain historical or archaeological evidence at its actual dating resolution and distinguish direct observations, retrospective accounts and later interpretations. Proposed or legally authorized works are not evidence that construction or ecological recovery occurred.

Exclude untraceable numerical claims from quantitative synthesis. Distinguish missing evidence from evidence of absence. Grade provenance, chronology, effective resolution, preservation, spatial coverage and data completeness separately. Published diagrams alone do not establish availability of complete assemblage matrices. Assess duplicate publications against the underlying core, survey or source to avoid counting the same evidence repeatedly.

## 7. Extraction and chronological comparability

The extraction unit is lake or sector × period × variable × method, linked to a unique underlying dataset or documentary source. Retain source location, coordinates and their precision, core/sample identifier, depth, count or measurement, unit, method, uncertainty and preservation flags. Archaeological and historical records additionally require site or feature, event versus document date, dating basis and range, spatial footprint, activity or intervention, evidential directness and alternative interpretation. Record hydraulic connections as documented, reconstructed or hypothetical.

Recover raw radionuclide activities, errors, density and independent age markers where possible. Evaluate radiocarbon material, reservoir effects and reworking. Apply multi-marker Bayesian models only where inputs justify them. Never use an event under test to date its assumed environmental response. Record uncertainty in parallel-core correlations. Compare effective sample spacing after age uncertainty and mixing, not nominal centimetres or interpolated grids. Do not interpolate across preservation gaps or hiatuses to fabricate continuous ecological histories.

## 8. Analytical decision rules

Recover complete diatom, pollen and macrofossil matrices with count sums and taxonomy. Compare them with compatible grain-size, organic-matter, geochemical and environmental records. Metal concentrations alone are not contaminant fluxes. Replicate comparisons across hydrogeomorphic zones where archives permit; a single core does not represent an entire lake.

Estimate rates of change under common temporal support, with sensitivity to bin width, smoothing, count standardisation and dissimilarity. Propagate chronological uncertainty through rates, GAM derivatives, change points, multivariate trajectories and lag comparisons. Use these methods only where data density and chronology support them; sparse evidence permits descriptive phases. A lag is unresolved when age uncertainty exceeds the apparent separation. Chronological alignment and independent process evidence are both needed for causal interpretation.

Compare archaeological or documentary changes with environmental trajectories at the coarsest defensible shared resolution. Test plausible competing pathways and source coverage. Neither a Roman-period label nor a modern engineering date establishes an ecological transition. Report negative and unresolved findings alongside supported changes.

## 9. Synthesis structure and outputs

Organize the synthesis as: Holocene landscape inheritance; reciprocal human–environment interactions during the last two millennia; the last 500 years as a bridge to modern engineering; recent ecological trajectories and rates; cross-lake mechanisms and restoration implications. Present observations, statistical support and interpretation separately.

Outputs comprise a provenance-rich catalogue and permitted data release, an archive/chronology methods paper, and a comparative synthesis with quantitative rate tests where qualified. The atlas will distinguish observed findings, proposed analyses and outstanding evidence. Restoration reference envelopes describe historical variability and potentially recoverable functions; current restoration outcomes require contemporary monitoring.

## 10. Feasibility and amendment record

At month 3, classify archives as quantitative candidates or contextual records. If one central lake lacks adequate data, retain a single-lake quantitative study with contextual comparison. If essential Smithsonian holdings are inaccessible, revise the institutional rationale before submission. The 24-month schedule governs the collections-led base design. The scientific core separates a proposed US$10,000 fellowship research-allowance request from a US$23,500 optional expansion for new cores, additional dating and multiproxy analyses, field trips, equipment and conferences. External support for the expansion is unconfirmed; these components require separately secured resources and permissions. The review does not imply a nationwide monitoring programme.

Amendment 28 September 2026: adds explicit reciprocal human–environment questions and historical-source extraction; establishes Burullus–Manzala priority and Mariut contingency; adopts four nested temporal windows; removes mandatory 1964–1965 breakpoint language; separates extended research design from completed evidence; and aligns outputs with the fellowship. Original documents remain in the archive. Subsequent changes to questions, eligibility, search cutoff or analytical rules require a dated record.