GeoSDG-XQ: A Quantum-Ready Explainable Geospatial Digital Twin for Indicator-Level Sustainable Development Intelligence

From Aquifers to Agenda 2030 - SDG 6 as the Case Demonstrator

FULL PAPER DRAFT - methodological/conceptual research paper

Kallol Saha

Founder-Director, Development Connects, Jharkhand, India | Geologist and Management Professional[Corresponding author email to be inserted]

Abstract

The Sustainable Development Goals (SDGs) are globally standardized but are still reported largely as statistical aggregates, limiting their usefulness for local decisions where development deficits are spatially heterogeneous, causally interconnected, and dynamically changing. This paper proposes GeoSDG-XQ, a quantum-ready, explainable geospatial intelligence architecture that converts SDG indicators into spatially addressable, time-aware and uncertainty-tagged decision objects. The method introduces an SDG Spatial Evidence Unit (SSEU) that binds an official indicator to geography, time, population, earth-system conditions, infrastructure, institutions, provenance, uncertainty, explanation and feasible intervention. Indicators are spatialized through seven classes: direct geocoding, administrative allocation, network representation, Earth-observation rasterization, model-derived surfaces, governance/institutional mapping and financial-flow mapping. A geospatial knowledge graph links these units across natural, social, infrastructural and institutional systems; GeoAI and geospatial foundation models support feature extraction and prediction; explainable and causal AI provide factor attribution, uncertainty and counterfactual intervention logic; and a geospatial digital twin enables historical reconstruction, current-state monitoring and future scenario simulation. SDG 6 is used as the proof-of-concept because water links geology, hydrogeology, infrastructure, ecosystems, behaviour, governance and finance. The proposed Water-SDG Twin integrates aquifers, rainfall, surface water, water quality, demand, service networks, wastewater, ecosystems and local institutions, while preserving the official definitions of all SDG 6 indicators. A local hydrological balance sheet and indicator-specific spatialization protocol are used to connect physical water accounts with service and governance outcomes. Finally, intervention selection is formulated as a constrained portfolio-optimization problem that can be solved classically and, for research experimentation, mapped to QUBO/QAOA workflows without claiming present quantum advantage. The framework is designed as an interoperable global methodology rather than a country-specific dashboard, with potential application to all SDGs and to investment prioritization by local governments, utilities, development agencies and climate-finance actors.

Keywords: GeoAI, SDG localization, explainable AI, geospatial digital twin, hydrogeology, water accounting, knowledge graph, quantum optimization, QAOA, SDG 6

Figure 1. GeoSDG-XQ architecture: converting official SDG indicators into spatial evidence, explanation, simulation and intervention optimization.

1. Introduction

The global Sustainable Development Goal framework is deliberately standardized: targets and indicators enable countries to report progress using common definitions. Yet the decisions that create or correct development outcomes are rarely made at the scale of a national statistic. Water failure occurs around particular aquifers, settlements, networks, institutions and seasons; sanitation gaps cluster around households and service systems; ecosystem loss occurs in identifiable reaches, wetlands and catchments; and governance failure is embedded in institutions with spatial jurisdictions. The resulting implementation gap is therefore partly a geographic problem: decision-makers need to know not only whether an indicator is improving, but where deficits are concentrated, what processes explain them, how conditions may evolve, and which intervention portfolio offers the greatest defensible improvement.

Goal 6 illustrates this problem sharply. The official framework covers safely managed drinking water, sanitation and hygiene, wastewater treatment, ambient water quality, water-use efficiency, water stress, integrated water-resources management, transboundary cooperation, water-related ecosystems, international cooperation and local community participation [1], [2]. These indicators span physical hydrology, hydrogeology, infrastructure, ecosystem condition, economics, finance, institutions and participation. A conventional GIS dashboard can display parts of this evidence, but display is not equivalent to a coherent methodological system for spatializing the indicator, explaining causal or contributory drivers, propagating uncertainty, simulating future states and comparing interventions.

This paper proposes GeoSDG-XQ, an extensible architecture for indicator-level sustainable-development intelligence. The core proposition is that every SDG indicator should be transformable into a spatially addressable, temporally indexed, provenance-preserving and uncertainty-tagged evidence object. The object can then participate in a geospatial knowledge graph, machine-learning workflows, explainability analysis, digital-twin simulations and constrained optimization. The approach is not intended to replace the United Nations indicator definitions or national statistical systems. Official values remain the reporting layer; GeoSDG-XQ adds a local analytical layer that makes those indicators decision-ready.

The framework is intentionally interdisciplinary. It combines an earth-system view associated with geology and hydrogeology, a spatial-computational view associated with GIS and GeoAI, and an institutional and investment view associated with management practice. This is important because SDG outcomes frequently emerge from interactions between biophysical limits and governance choices. The SDG 6 case therefore begins from the aquifer and watershed but extends to service delivery, institutions, financial flows and community participation.

2. Practice lineage and research gap

The proposed framework has a practical lineage in decentralized water governance. The Jharkhand water and sanitation experience documented in Abua Raj-Nawa Khoij emphasized the need to empower Panchayati Raj Institutions and Village Water and Sanitation Committees, link infrastructure with community management, promote groundwater recharge and conjunctive use, and modernize programme delivery [3]. The same programme history documented GPS-supported mapping of water sources and service inequalities, contaminant-risk mapping, mobile and call-centre enabled monitoring, and real-time monitoring systems. Its future perspective explicitly called for context-specific R&D, institutional convergence and expanded real-time monitoring. GeoSDG-XQ updates this practice logic for the era of Earth observation, geospatial foundation models, knowledge graphs, explainable AI, digital twins and hybrid optimization.

Recent GeoAI research shows rapid expansion in the use of machine learning with spatiotemporal data, while also identifying interpretability, bias, reproducibility and transferability as persistent challenges [4]. Hydrological AI literature similarly highlights strong predictive capability for nonlinear processes alongside the need for explainability and transfer learning [5]. Meanwhile, digital twins are increasingly used in water infrastructure and river-basin contexts to integrate real-time observations, models, scenarios and operational decisions [6], [7]. These advances are important but are usually organized around a technical system, a sectoral application or a model. They are not yet, by themselves, a general protocol for spatializing every SDG indicator while preserving official semantics and connecting physical, social, institutional and financial evidence.

The research gap addressed here is therefore methodological: how can a single geospatial architecture transform heterogeneous SDG indicators into comparable spatial evidence units while maintaining official indicator integrity, explicit uncertainty, explainability and decision traceability? A second gap concerns action: how can the same architecture move from measurement to scenario simulation and intervention prioritization without presenting opaque AI recommendations as unquestionable answers?

3. Research objectives and contributions

GeoSDG-XQ has five research objectives:

· Define a universal SDG Indicator Spatialisation Protocol that preserves official indicator semantics while representing relevant evidence at decision scales.

· Create the SDG Spatial Evidence Unit (SSEU) as the atomic object linking indicator, location, time, population, earth-system state, infrastructure, institution, provenance, uncertainty, explanation and intervention.

· Integrate GeoAI, spatial knowledge graphs and explainable/causal AI within a time-aware geospatial digital twin.

· Demonstrate the methodology through SDG 6 using a local hydrological balance sheet and an aquifer-to-governance systems model.

· Formulate intervention planning as a transparent constrained optimization problem with a classical baseline and a quantum-ready QUBO/QAOA research pathway.

4. The SDG Spatial Evidence Unit (SSEU)

The methodological centre of GeoSDG-XQ is the SDG Spatial Evidence Unit. For an indicator i, an SSEU is represented as:

SSEUᵢ = {I, G, T, P, E, N, Inst, V, Prov, U, X, A}

where I is the official indicator identifier and definition; G is spatial geometry or spatial support; T is time; P is the relevant population or denominator; E is earth-system state; N is service or infrastructure network state; Inst is the responsible or participating institution; V is the observed or modelled value; Prov is provenance; U is uncertainty; X is the explanation vector; and A is a set of feasible actions or interventions. Not every indicator needs every field, but the schema forces the analytical system to record what is known, what is inferred, and what is absent.

The SSEU deliberately separates official measurement from local analytical evidence. A model-derived village-scale water-stress surface, for example, may be valuable for planning, but it must not be mislabelled as the official national value of indicator 6.4.2. GeoSDG-XQ therefore maintains a semantic flag for each value: official, administratively disaggregated, directly observed, remotely sensed, model-estimated or scenario-generated.

5. SDG Indicator Spatialisation Protocol

Class	Method	Typical representation
S1 Direct geocoding	Indicator already tied to points/households/assets	water source, treatment plant, household service
S2 Administrative allocation	Statistic reported for a jurisdiction	district/local-body indicator with known denominator
S3 Network representation	Outcome depends on flows or connectivity	water distribution, wastewater collection, transport access
S4 Earth-observation rasterization	Indicator or driver is observable from imagery	surface water, wetland extent, land cover, settlement
S5 Model-derived surface	Indicator requires inference from physical/social models	groundwater stress, exposure, future demand
S6 Governance/institutional mapping	Indicator concerns policy, participation or institutional performance	IWRM, community participation
S7 Financial-flow mapping	Indicator concerns finance, investment or programme support	ODA, project expenditure, capex allocation

Spatialisation follows five rules: (1) preserve the official numerator, denominator and conceptual definition; (2) identify the spatial support appropriate to the underlying phenomenon rather than forcing every indicator into administrative polygons; (3) document transformations and uncertainty; (4) distinguish observation from proxy and prediction; and (5) allow multi-scale nesting so that grid, asset, settlement, local body, basin and national representations can coexist.

6. GeoAI, knowledge graphs and explainability

GeoAI provides the computational layer for extracting features from Earth observation, detecting assets, modelling spatiotemporal relationships and estimating values where direct observation is incomplete. Current ArcGIS development explicitly includes geospatial foundation models and embeddings that can be adapted across mapping, imagery and predictive workflows [8]. In GeoSDG-XQ these models are treated as reusable feature generators rather than as autonomous policy engines.

A spatial knowledge graph connects the heterogeneous evidence. For SDG 6, entities may include Aquifer, Well, RiverReach, HouseholdCluster, TreatmentPlant, Pipeline, Panchayat, Municipality, WaterCommittee, Project, Budget, Indicator and Intervention. Relations can include RECHARGES, SUPPLIES, SERVES, DISCHARGES_TO, MANAGED_BY, FINANCED_BY, AFFECTS and CONTRIBUTES_TO. ArcGIS Knowledge demonstrates how spatial and non-spatial enterprise data can be fused into graph structures on a shared ontology [10]. The graph makes cross-sector dependencies explicit and enables traceable queries such as: which settlements depend on wells intersecting a declining aquifer and are governed by institutions with low maintenance capacity?

Explainability is required at three levels. First, predictive explanation identifies which variables contributed to an estimate using methods such as SHAP-type feature attribution, partial dependence or interpretable surrogate models. Second, spatial explanation tests whether the contribution is locally stable or changes by geography. Third, decision explanation communicates the chain from evidence to recommended intervention, including uncertainty, constraints and counterfactual alternatives. Where causal identification is not justified, the system uses the terms contribution or association rather than causation.

7. The geospatial digital twin

A geospatial digital twin is more than a 3D visualization. It is a time-aware representation that connects physical objects, processes and relationships and can represent historical, current and future states [9]. GeoSDG-XQ uses the twin as the temporal execution environment for the SSEU and knowledge graph. Its state vector is updated from observations and models; scenario engines perturb drivers such as rainfall, population, infrastructure performance or policy; and indicator engines recompute expected SDG outcomes. River-basin digital-twin research similarly emphasizes real-time monitoring, historical observations, predictive modelling, high-performance computing and what-if analysis [7].

The twin has four operating modes: observe (current indicator evidence), diagnose (explain deficits), simulate (test future or intervention scenarios) and optimize (select portfolios under constraints). This separation is important for public-sector accountability because it distinguishes what the system has measured from what it predicts and from what it recommends.

8. SDG 6 case demonstrator: the Water-SDG Twin

SDG 6 is selected because it creates a demanding test of the architecture. It requires simultaneous representation of household services, groundwater and surface-water systems, water quality, wastewater, ecosystems, economic efficiency, institutional integration, finance and community participation [1], [2]. The proposed Water-SDG Twin therefore uses six interacting layers:

· Subsurface earth layer: lithology, structures, geomorphology, aquifer geometry, weathered/fractured zones, groundwater heads, recharge, storage, groundwater quality and contaminant pathways.

· Surface hydrology-climate layer: precipitation, evapotranspiration, runoff, rivers, reservoirs, wetlands, soil moisture, drought/flood conditions and climate scenarios.

· Infrastructure layer: wells, intakes, pumps, reservoirs, treatment facilities, pipelines, household connections, sanitation systems, sewers, wastewater plants, meters and recharge structures.

· People and demand layer: population, settlements, domestic demand, agriculture, livestock, industry, institutions and future demand scenarios.

· Economy and finance layer: CAPEX, OPEX, tariffs, water productivity, programme expenditure, investment gaps and intervention costs.

· Governance and community layer: local bodies, utilities, water committees, user organizations, plans, participation, grievances, maintenance arrangements and programme convergence.

8.1 Local hydrological balance sheet

For each selected spatial accounting unit - preferably both a hydrologically meaningful unit and the local administrative unit that makes decisions - the twin maintains a water balance:

ΔS = P + Qin + Gin + R + Import + Reuse - ET - Qout - Gout - W - EFR - Loss

where ΔS is change in storage; P precipitation contribution; Qin/Gin surface and groundwater inflow; R recharge; Import externally supplied water; Reuse treated reuse; ET evapotranspiration; Qout/Gout outflow; W withdrawals; EFR environmental-flow requirement; and Loss measured or estimated system losses. The balance can be maintained separately for surface water, shallow groundwater, deeper aquifers, harvested rainwater and treated wastewater, depending on local data availability.

The hydrological balance is an analytical account, not a substitute for the official SDG calculation. For indicator 6.4.2, official definitions and nationally reported parameters remain authoritative; local accounts are used to explain spatial stress and test interventions.

8.2 Indicator-wise spatialisation of SDG 6

Indicator	Official focus	GeoSDG-XQ spatial expression	Class
6.1.1	Safely managed drinking water	household/settlement service surface; source-to-user network; availability/quality layers	S1/S2/S3
6.2.1	Sanitation and hygiene	household/service areas; sanitation chain; handwashing access	S1/S2/S3
6.3.1	Wastewater safely treated	generation -> collection -> treatment -> discharge/reuse flow graph	S3/S5
6.3.2	Ambient water quality	sample points + river/lake/aquifer interpolation and risk surfaces	S1/S4/S5
6.4.1	Water-use efficiency	sectoral withdrawals/consumption linked to value added by zone	S2/S5/S7
6.4.2	Water stress	resource-availability and withdrawal surfaces with EFR constraints	S5
6.5.1	IWRM implementation	institutional jurisdiction and component-wise governance score maps	S2/S6
6.5.2	Transboundary cooperation	shared basin/aquifer graph and operational-arrangement coverage	S3/S6
6.6.1	Water-related ecosystems	EO-derived extent/change of wetlands, rivers, lakes and linked aquifer/ecosystem zones	S4/S5
6.a.1	International cooperation	project/finance geographies linked to outputs and indicator change	S7
6.b.1	Local community participation	local administrative units, committees, rules, participation and functionality evidence	S2/S6

8.3 Explainable water-stress diagnosis

A typical model output should not be a single risk class. The diagnostic object contains: predicted or measured condition; confidence interval or probability; principal contributing factors; spatial context; data lineage; and counterfactual interventions. For example, a high local water-stress classification may be associated with seasonal aquifer drawdown, irrigation abstraction, low recharge, distribution loss and population growth. The interface should present factor contributions and then test whether proposed actions - recharge, demand management, leakage reduction, wastewater reuse, cropping changes or source diversification - materially change the indicator trajectory. The system must avoid presenting statistical attribution as causal proof unless the design supports causal inference.

9. State-of-the-art sensing and data architecture

The proposed architecture is technology-agnostic but designed to use state-of-the-art inputs where they add measurable value. Earth observation can include multispectral, SAR and thermal data; UAV and LiDAR can support high-resolution asset and terrain mapping; GNSS/NavIC can improve field geolocation; IoT sensors and smart meters can update flows and quality; administrative MIS and financial systems provide service and programme evidence; mobile/citizen reporting can contribute localized observations; and geospatial foundation models can accelerate feature extraction and similarity search [8]. The key methodological requirement is provenance: every derived value must retain source, date, processing chain, spatial resolution and uncertainty.

10. Quantum-ready intervention optimization

Quantum computing is included as a research horizon rather than a marketing claim. Many practical planning problems in SDG implementation are combinatorial: selecting a limited set of projects from thousands of candidates, siting sensors or treatment assets, allocating scarce water among users, designing networks, and assembling portfolios that improve several SDGs while respecting budget, equity, environmental and engineering constraints. These problems can first be formulated and solved with mature classical methods.

For selected binary formulations, the same decision problem can be mapped to Quadratic Unconstrained Binary Optimization (QUBO). IBM's QAOA documentation describes QAOA as a hybrid quantum-classical iterative method for combinatorial optimization [13], while warm-start approaches can incorporate a classical relaxation before quantum search [14]. GeoSDG-XQ therefore adopts a benchmark-first protocol: establish a classical optimum or strong heuristic baseline; formulate the identical objective and constraints as QUBO where feasible; test QAOA on small/synthetic instances; and report solution quality, runtime, scaling and robustness without claiming quantum advantage unless demonstrated.

A generic intervention-portfolio objective can be written as:

maximize  Σᵢ wᵢ·ΔSDGᵢ(x)  -  λ·Cost(x)  -  μ·Risk(x)  -  ν·Inequity(x)

subject to budget, engineering feasibility, ecological thresholds, minimum service obligations and institutional constraints. The weights are not hidden AI parameters: they are explicit policy choices that can be tested through sensitivity analysis and participatory decision processes.

11. Prototype implementation in the ArcGIS ecosystem

The proposed prototype can be implemented using an interoperable ArcGIS-centred stack without locking the scientific method to a single vendor. ArcGIS Pro/Enterprise can manage authoritative spatial layers; geospatial foundation models can support GeoAI workflows [8]; ArcGIS Knowledge can represent entity-relationship structures and provenance [10]; dashboards and web applications can communicate indicator evidence; and digital-twin patterns can support time-aware scenario visualization [9]. The IIT Tirupati Navavishkar I-Hub Geo-Intel Lab explicitly identifies real-time spatial monitoring, digital twins, GeoAI, decision systems for local governments and public-good applications as focus areas [11]. Its Center of Competence with Esri India is intended to advance research and industry-academia collaboration in GIS, remote sensing, AI/ML and geospatial technologies [12]. This makes the proposed methodology a plausible research-collaboration topic rather than merely a software demonstration.

12. Validation and research design

A publishable prototype should be validated in stages rather than presenting conceptual diagrams as completed evidence. The recommended research design is:

· Stage 1 - Ontology validation: expert review of SSEU fields, indicator semantics and spatialisation classes against official UN metadata.

· Stage 2 - Data-pipeline validation: reproducibility tests for provenance, coordinate systems, temporal alignment, missing data and uncertainty propagation.

· Stage 3 - SDG 6 pilot: one rural local body and one urban local body, preferably nested within hydrologically meaningful catchment/aquifer units.

· Stage 4 - Model validation: spatial cross-validation, calibration, uncertainty intervals and comparison with transparent baseline models.

· Stage 5 - Explainability validation: stability and fidelity of explanations; expert and community review of whether explanations match field reality.

· Stage 6 - Scenario validation: hindcasting where historical data permit and sensitivity tests for climate, demand and infrastructure assumptions.

· Stage 7 - Optimization validation: compare recommended portfolios with classical optimization baselines; quantum experiments are reported as research benchmarks.

· Stage 8 - Transferability test: reproduce the workflow in a second geography with different geology, governance and data availability.

Potential performance metrics include indicator-estimation error, spatial calibration, uncertainty coverage, explanation stability, reproducibility, computational cost, intervention cost-effectiveness and decision usefulness. Qualitative governance metrics should be explicitly separated from physical-model accuracy.

13. Global transferability

GeoSDG-XQ is designed as a method, not an India-only database. The global layer uses official SDG indicator identifiers and metadata; national adapters map country data standards; subnational adapters map administrative and hydrological units; and local implementations add assets, institutions and community evidence. A country with sophisticated sensors may operate near real time, while a data-constrained setting may rely more heavily on remote sensing, periodic surveys and modelled uncertainty. The architecture remains the same because the SSEU records both value and evidence quality.

The same method can extend beyond SDG 6. Examples include SDG 2 using crop, soil, market and livelihood evidence; SDG 3 using facility access, exposure and disease surfaces; SDG 5 using service, tenure and institutional evidence; SDG 11 using settlement, transport, risk and infrastructure networks; SDG 13 using hazard, vulnerability and adaptation investments; and SDG 15 using ecosystem condition, fragmentation and restoration. The key is not to force every indicator into an artificial map but to assign the correct spatial representation class and record uncertainty.

14. Governance, ethics and safeguards

A system that localizes development indicators can also create risks: false precision, privacy intrusion, biased proxies, exclusion of unconnected populations, automation of contested policy priorities and overconfidence in model outputs. GeoSDG-XQ therefore requires data minimization for household-level records, role-based access, aggregation or privacy-preserving release where needed, transparent model cards, audit trails, community validation of participatory indicators, and explicit uncertainty. Decisions with distributional consequences should remain human decisions supported by the system, not automated outcomes.

15. Discussion

The principal contribution of GeoSDG-XQ is not any one technology. Its novelty lies in the methodological coupling of official SDG semantics, spatial evidence, earth-system understanding, institutional mapping, explainability, digital-twin simulation and transparent optimization. This coupling is especially relevant for sustainability because physical resource limits and management decisions operate on different spatial and temporal scales.

The SDG 6 demonstrator makes the role of geological reasoning explicit. Aquifers are three-dimensional, heterogeneous and only partially observed; administrative boundaries rarely coincide with recharge areas or groundwater flow systems. A useful SDG intelligence system must therefore allow an administrative decision unit to inherit evidence from the hydrological systems on which it depends. This is a deeper proposition than simply colouring administrative polygons by indicator value.

The management dimension is equally important. Infrastructure may exist yet fail to provide a service because operations, tariffs, maintenance, institutional responsibility or community participation are weak. The SSEU and knowledge graph provide a way to represent these dependencies rather than treating governance as explanatory text outside the GIS.

The quantum-ready component should remain experimentally modest. The immediate value of GeoSDG-XQ comes from better data organization, spatial reasoning, explainability and scenario analysis. Quantum optimization becomes relevant only for well-defined combinatorial subproblems where an identical classical baseline exists.

16. Proposed collaboration and development roadmap

Phase	Indicative horizon	Output
Phase 0 | Conference	Sep-Nov 2026	method paper, architecture, SDG 6 indicator ontology, demonstrator design
Phase 1 | Minimum viable twin	0-6 months	rural + urban pilot; hydrological balance; selected 6.1.1/6.3.2/6.4.2/6.6.1/6.b.1 layers
Phase 2 | Full SDG 6	6-12 months	all 11 indicators; knowledge graph; XAI; scenario engine
Phase 3 | Multi-SDG	12-24 months	cross-SDG dependencies and investment portfolio optimizer
Phase 4 | Global adapters	18-36 months	second-country validation; open ontology/API specification
Research track | Quantum	parallel	QUBO formulation and QAOA benchmarking for intervention selection and siting

17. Conclusion

GeoSDG-XQ proposes a shift from static SDG mapping to explainable spatial decision intelligence. Its basic unit, the SSEU, binds official indicator semantics to place, time, earth-system state, infrastructure, institutions, provenance, uncertainty, explanation and action. The SDG 6 case demonstrates how geology and hydrogeology can be connected to service delivery, ecosystems, governance and finance through a Water-SDG Twin and local hydrological balance sheet. GeoAI and foundation models accelerate extraction and prediction; knowledge graphs represent dependencies; explainable AI communicates drivers and uncertainty; the digital twin tests future states; and classical/quantum-ready optimization compares intervention portfolios. The immediate scientific task is not to claim a finished global platform, but to validate this architecture rigorously in contrasting rural and urban pilots and then test transferability. If successful, the approach can provide a methodological bridge from Agenda 2030 reporting to locally actionable, transparent and globally interoperable sustainable-development intelligence.

Acknowledgement of practice basis

The conceptual development draws on lessons from water and sanitation programme reform, community institutions, GIS-enabled monitoring and real-time programme management documented in Jharkhand during 2010-2014 [3]. The author was acknowledged in that publication for substantive contribution to content and creative presentation during his association with the State Water and Sanitation Mission.

References

[1] United Nations Department of Economic and Social Affairs (UN DESA), “Goal 6: Ensure availability and sustainable management of water and sanitation for all,” 2026. https://sdgs.un.org/goals/goal6

[2] United Nations Statistics Division (UNSD), “SDG Indicators Metadata Repository,” updated 24 August 2026. https://unstats.un.org/sdgs/metadata/

[3] S. Prasad, Abua Raj-Nawa Khoij: From Policy to Practice - Water and Sanitation in Jharkhand. Drinking Water and Sanitation Department, Government of Jharkhand, with UNDP support, 2016.

[4] S. Wang et al., “Mapping the landscape and roadmap of geospatial artificial intelligence (GeoAI) in quantitative human geography: An extensive systematic review,” International Journal of Applied Earth Observation and Geoinformation, vol. 128, 103734, 2024. doi:10.1016/j.jag.2024.103734.

[5] R. Maity, A. Srivastava, S. Sarkar, and M. I. Khan, “Revolutionizing the future of hydrological science: Impact of machine learning and deep learning amidst emerging explainable AI and transfer learning,” Applied Computing and Geosciences, vol. 24, 100206, 2024. doi:10.1016/j.acags.2024.100206.

[6] P. Ghorbani Bam et al., “Digital Twin Applications in the Water Sector: A Review,” Water, vol. 17, no. 20, 2957, 2025. doi:10.3390/w17202957.

[7] D. Pal et al., “Blueprint conceptualization for a river basin’s digital twin,” Hydrology Research, vol. 56, no. 3, pp. 197-212, 2025. doi:10.2166/nh.2025.111.

[8] Esri, “Introducing Geospatial Foundation Models in ArcGIS,” 12 July 2026. https://www.esri.com/arcgis-blog/products/arcgis-pro/geoai/introducing-geospatial-foundation-models-in-arcgis

[9] Esri, “Digital Twin Technology & GIS - What is a Digital Twin?” accessed 19 September 2026. https://www.esri.com/en-us/digital-twin/overview

[10] Esri, “What’s new in ArcGIS Knowledge 12.1 (2026 Q2),” 21 May 2026. https://www.esri.com/arcgis-blog/products/arcgis-knowledge/analytics/whats-new-in-arcgis-knowledge-12-1

[11] IIT Tirupati Navavishkar I-Hub Foundation, Geo-Intel Lab, “Spatial Intelligence: Driving Public Good through Spatial Innovation,” accessed 19 September 2026. https://geo.intel.iittnif.com/

[12] IIT Tirupati Navavishkar I-Hub Foundation, “Center for Competence in Collaboration with ESRI,” accessed 19 September 2026. https://geo.intel.iittnif.com/centers-and-labs/center-for-competence-esri

[13] IBM Quantum, “Quantum approximate optimization algorithm,” accessed 19 September 2026. https://quantum.cloud.ibm.com/docs/en/tutorials/quantum-approximate-optimization-algorithm

[14] IBM Quantum, “Warm-start QAOA with the Optimization Mapper Qiskit addon,” accessed 19 September 2026. https://quantum.cloud.ibm.com/docs/en/tutorials/warm-start-qaoa