An Interactive Platform for Design Hydrograph Estimation in Small and Ungauged Basins: Pilot Implementation in the Lazio Region, Italy

Abstract

Estimating design hydrographs in small and ungauged basins remains a significant challenge, primarily due to limited hydrometeorological data and the operational complexity of advanced modelling tools. This study presents an interactive digital twin platform to support hydrological modelling in such contexts. The aim of the proposed platform is to integrate three hydrological models—EBA4SUB (event-based rainfall–runoff model), COSMO4SUB (continuous rainfall–runoff model), and Virtual Rain (stochastic rainfall generator)—and automates key pre-processing tasks, including watershed delineation, Curve Number estimation, and rainfall input generation. Built on a three-tier architecture, the system comprises an interactive front end, a back-end database with spatial and meteorological data, and a suite of computational routines developed in Python and C#. The platform was deployed across the Lazio Region (Italy) for basins with contributing areas smaller than 400 km². Users can interactively select watersheds via a map-based interface, obtain preliminary hydrological characterizations, and export model-ready inputs and outputs. The proposed platform offers several advantages: it reduces model preparation time, facilitates access to advanced modelling tools, standardizes input data at the regional level, and ensures reproducible pre-processing workflows. By lowering the technical and time barriers of hydrological modelling, the digital twin provides an effective framework for bringing science-based tools closer to real-world practice.

1. Introduction

Designing reliable hydrological estimates for small and ungauged basins has long represented a persistent challenge in both academic research and professional practice [1,2]. These basins, often characterized by limited or nonexistent streamflow measurements, pose significant obstacles to effective planning and risk assessment, particularly in estimating key design parameters such as peak discharge and complete runoff hydrographs. This limitation has important implications for flood risk management, stormwater infrastructure sizing, and land use planning, where accurate design flows are essential to protect communities and ecosystems [3,4,5].

The core challenge lies in data scarcity. Most small basins lack the continuous streamflow and rainfall records required to calibrate advanced hydrological models. As a result, models that depend on parameter fitting, spatial variability of inputs, or long-term time series are difficult to apply in such contexts. This scarcity of data precludes the routine use of distributed, physically based, or conceptual models in many professional settings, pushing practitioners toward simpler, though less rigorous, approaches [6,7,8].

Historically, one such alternative has been the Rational Formula, a method developed in the 19th century and still widely used today. Its enduring popularity stems largely from its simplicity, minimal input requirements, and ease of application—factors particularly attractive to engineers and planners working with limited data, time, or technical resources [9,10,11]. The Rational Formula estimates peak discharge from rainfall intensity, runoff coefficients, and drainage area, making it especially suitable for small catchments where real-time data collection and model calibration are impractical. Nevertheless, the method has well-documented scientific limitations, including its inability to simulate hydrograph shapes, internal runoff processes, or time-varying inputs. Despite these shortcomings, its ubiquity persists, often at the expense of more robust hydrological analyses [12,13,14].

In response, the hydrological modelling community has developed a range of alternative methods aimed at balancing scientific robustness with operational feasibility. Some notable examples are HEC-HMS [15], SWAT [16,17], MIKE SHE [18], TOPMODEL [19]. These methods—spanning from event-based to continuous simulation models—are designed to yield more realistic runoff estimates while still operating with limited data. A key advantage of these approaches is their ability to incorporate widely available spatial and meteorological datasets, such as Digital Elevation Models (DEMs), land use maps, and design rainfall hyetographs, which are often accessible even in ungauged settings [20,21,22]. In some cases, stochastic rainfall generators or remote sensing inputs can also be used to simulate synthetic rainfall events for hydrological modelling.

Despite these theoretical advantages, the adoption of modern tools by practitioners remains limited. A primary barrier is the perceived technical complexity of their implementation. Unlike the Rational Formula, event-based and continuous models typically require users to engage with more advanced hydrological concepts, such as DEM preprocessing, catchment discretization, excess rainfall estimation (e.g., Curve Number or Green-Ampt infiltration), and synthetic rainfall simulation. Many practitioners lack the time, specialized training, or software skills needed to perform these preparatory steps, particularly in fast-paced consultancy or municipal engineering environments [23,24]. Moreover, these models often involve multiple pre-processing tools, file format conversions, and GIS tasks, which collectively increase the time and effort required for their application—even for relatively simple runoff events.

This disconnect between scientific methodology and real-world usability has led to a paradox: although more accurate models exist, their limited usability has prevented them from displacing outdated and less accurate methods. To bridge this gap, there is a clear need for tools that preserve the theoretical strengths of modern hydrological models while lowering the entry barrier for their application in professional settings.

In this context, we propose an interactive platform specifically designed to support hydrograph estimation in small and ungauged basins. The platform directly addresses the primary obstacles to using advanced hydrological models by automating the most time-consuming and technically demanding steps of the modelling workflow. These include DEM preprocessing, rainfall scenario generation, catchment delineation, and input file formatting. Importantly, the platform is not intended to function as a “black box” but rather an interface that facilitates users to autonomously perform the modelling.

To the best of our knowledge, there are no web-service–based digital platforms available that perform automatic pre-processing of inputs for the application of hydrological models. The platforms, like HydroDS [25], PAVICS-Hydro [26], ATHIS [27] offer useful web-based tools, but they share common limitations such as dependence on available data and the need for particular technical expertise.

The platform integrates three hydrological models developed for small and ungauged catchments: EBA4SUB (event-based rainfall–runoff model), COSMO4SUB (continuous rainfall–runoff model), and Virtual Rain (stochastic rainfall generator). Each balances robustness with parsimony, and together they provide a suite of modelling options for both event-based and continuous simulation contexts. The platform offers a streamlined and user-friendly interface, enabling practitioners to efficiently prepare inputs and perform simulations without extensive GIS or programming expertise.

While the literature in recent years has proposed various advanced models [15,16,17,18,19], few studies have focused on facilitating their practical implementation by reducing the technical burden on users. This paper seeks to address this implementation gap by introducing a prototype tool that combines scientific rigour with operational feasibility. The proposed interactive platform serves as a missing link between model development and real-world application, thus encouraging broader uptake of advanced hydrological modelling approaches in data-scarce contexts.

The manuscript is structured as follows: Section 2 provides a theoretical overview of the three models integrated into the platform. Section 3 examines the conventional workflow commonly followed by practitioners, along with its limitations. Section 4 presents the platform in detail, highlighting its potential benefits in terms of usability, reproducibility, and computational efficiency.

2. Hydrological Modelling for Small and Ungauged Basins

Over the past decade, several modelling improvements and adaptations in both event-based and continuous hydrological modelling approaches were proposed [28,29] with particular emphasis on contexts characterized by limited or no observational data—settings where the parameter calibration is not possible and the Rational Formula is traditionally applied. In this work, we refer to two rainfall–runoff models specifically designed for small and ungauged basins: EBA4SUB and COSMO4SUB.

The Event-Based Approach for Small and Ungauged Basins (EBA4SUB) is an empirical, conceptual, lumped hydrological model designed to simulate the transformation of rainfall into runoff in small (contributing area smaller than a few hundred km²), data-scarce catchments [30,31]. It is a simplified model that excludes slower hydrological processes such as baseflow, evaporation, and snowmelt, focusing solely on surface runoff.

EBA4SUB consists of three main components: (1) gross rainfall identification, (2) excess rainfall estimation, and (3) excess rainfall–runoff transformation. In component 1, the synthetic design hyetograph is derived from Depth–Duration–Frequency (DDF) curves, assuming predefined distributions (e.g., Chicago-type, triangular, rectangular) and rainfall critical duration (time of concentration). In component 2, excess rainfall is estimated via the Curve Number for Green–Ampt (CN4GA) method, which combines the Curve Number (CN) approach [32] with the Green–Ampt infiltration equation [33]. The CN4GA method [34] first applies the CN approach to estimate ponding time and cumulative hyetograph excess rainfall, then refines infiltration dynamics within the hyetograph using the Green–Ampt equation, blending empirical efficiency with improved process representation. Component 3 transforms excess rainfall into runoff using the Width-Function Based Instantaneous Unit Hydrograph (WFIUH), derived from a hydrologically corrected Digital Elevation Model (DEM) [35]. The WFIUH accounts for both hillslope and channel travel times, with surface velocities estimated empirically from slope and land cover data, while channel velocities are calibrated to align the hydrograph centre of mass with the basin lag time (set to 60% of the time of concentration, in line with [35]).

Based on these components, EBA4SUB automatically and empirically quantifies most required input parameters, such as the Curve Number and concentration time. The only required inputs are the Depth–Duration–Frequency (DDF) curve parameters, the hydrologically corrected DEM, and land cover data for the watershed.

COSMO4SUB (Continuous Simulation Model for Small and Ungauged Basins) [20,36] is a continuous, empirical, conceptual lumped rainfall–runoff model composed of three modules, following the same conceptual framework as EBA4SUB: (1) rainfall simulation, (2) excess rainfall estimation, and (3) transformation of excess rainfall into runoff. The main difference from EBA4SUB is that it does not rely on a design hyetograph but instead uses long synthetic rainfall time series as input. This allows continuous runoff simulation by dynamically accounting for antecedent soil moisture, estimated from the cumulative rainfall over the five days preceding each event. Grimaldi et al. [20] illustrate and discuss the benefits of this continuous approach compared to event-based methods.

Module 1 operates on sub-daily precipitation time series, sourced either from observations (e.g., rain gauge stations) or synthetic series generated by tools such as Virtual Rain. This framework produces high-resolution rainfall inputs by combining a daily rainfall generator [37,38] with a multifractal sub-daily disaggregation model [39]. Its main advantage is the minimal data requirement: calibration can be performed using only a daily rainfall series and DDF parameters. Unlike most existing rainfall simulation tools, it does not require sub-daily observed data for calibration, making it highly suitable for data-scarce regions.

Module 2 estimates excess rainfall using a continuous version of CN4GA. Unlike EBA4SUB, COSMO4SUB introduces an additional parameter—the rainfall event separation time (Ts)—which defines the minimum dry interval required between two storms for initial abstraction to reset. The model also evaluates antecedent moisture conditions (AMC) using rainfall over the preceding five days, adjusting the CN-II value (average soil moisture) toward CN-I (dry) or CN-III (wet), consistent with the original Curve Number methodology.

Module 3 transforms excess rainfall into runoff using the WFIUH, following the same approach as in EBA4SUB.

By supporting long-term streamflow simulations, COSMO4SUB enables the estimation of annual maximum peak discharges, the derivation of peak discharge–return period relationships, and the generation of extreme hydrograph ensembles. Importantly, it requires the same input data as EBA4SUB.

To encourage adoption of these models in professional practice, substantial resources were invested in practitioner engagement and training. However, uptake proved limited. Despite widespread recognition of the Rational Method’s limitations, it remained the preferred choice, primarily due to its simplicity and ease of use. This feedback provided a strong impetus for developing the digital twin platform—presented in the following sections—designed to reduce perceived complexity and facilitate broader use of advanced hydrological models.

3. The Common Practitioner Workflow

To apply the previously described hydrological models, practitioners must follow a streamlined and reproducible workflow. The initial preparatory step for analyzing the hydrological response of an ungauged small catchment involves preprocessing the DEM to remove depressions (pits), where surface runoff would otherwise accumulate, and flat areas, where flow direction is ambiguous. Algorithms such as that proposed by Wang and Liu [40] are commonly used for this purpose. The hydrographic network derived from official technical cartography, provided by District Hydraulic Authorities or similar agencies, must then be imposed onto the DEM. From the selected outlet section, catchment boundaries are delineated using approaches such as Wang and Liu [40]. Once defined within the reference coordinate system and DEM spatial resolution, land use and land cover maps [41] and Curve Number maps [42] can be clipped to ensure spatial alignment and resolution consistency.

This workflow enables the derivation of hydrological parameters required for WFIUH computation. Specifically, slopes along the eight cardinal directions are calculated for each cell; hillslope inclination is defined as the mean of these values, while flow direction (FD) is assigned to the direction of maximum slope. The FD map allows computation of Flow Accumulation (FA), from which the hydrographic network is extracted by applying a user-defined threshold. Flow velocities are then assigned to both hillslope and channel cells based on slope and land cover characteristics. Finally, travel times of precipitation contributions from each cell to the outlet are computed, distinguishing between hillslope and channel pathways. An initial distribution of travel times is generated and subsequently adjusted by calibrating channel flow velocities so that the lag time corresponds to 60% of the time of concentration.

From a practitioner’s perspective, completing this procedure manually demands considerable effort: even an experienced GIS analyst with appropriate tools typically requires between one and five working days, depending on catchment size and complexity. Moreover, the workflow involves several subjective choices—such as pit filling, flow-direction assignment, network burning, and spatial averaging of parameters—that can introduce variability and affect the results. This complexity highlights the potential value of a digital twin, which can automate repetitive steps, standardize subjective decisions, and reduce both the time and uncertainty associated with conventional GIS-based hydrological preprocessing.

4. The Proposed Interactive Platform

The objective is to translate the procedure outlined in the preceding sections into a practical and time-efficient workflow, designed to be executed within a few minutes. In addition to ensuring rapid implementation, the workflow provides preliminary diagnostic outputs that support users in navigating the modelling framework and in refining the application of the chosen hydrological model. From an operational standpoint, the digital twin framework is conceived as a structured process articulated in three main phases (Figure 1).

• Preliminary Analysis. The user engages with the platform via a geospatial interface to delineate the area of interest, select the watershed, and retrieve an initial hydrological and morphological characterization (e.g., DDF parameters for selected return periods, watershed area, concentration time, and Curve Number).
• Event-Based Modelling. Once the watershed is confirmed, the user is directed to the EBA4SUB module, which supports detailed event-based hydrological simulations and scenario refinement.
• Continuous Modelling. For applications requiring continuous simulation, the user selects a representative rain gauge and proceeds either to the Virtual Rain module for stochastic rainfall generation or directly to the COSMO4SUB model for extended hydrological analysis.

From a system architecture perspective, the proposed digital twin comprises three core operational environments (Figure 2):

• Interactive Environment. The front-end interface facilitates user interaction through a navigable map overlaid with the drainage network and optional rain gauge locations. Based on user input, contextual windows are dynamically generated to present preliminary analysis results.
• Back-End and Database. These components store and manage pre-processed geospatial and hydrological datasets, including a hydrologically corrected DEM, flow direction and accumulation maps, a discretized drainage network, land use classifications, and regionalized Depth–Duration–Frequency (DDF) equations. The back-end executes all computational algorithms, while the database preserves input, output, and intermediate results.
• Computational Routines. This environment orchestrates model execution by processing user inputs, invoking the appropriate hydrological modules, and managing data exchange with the back-end and database.

Regarding the Interactive Environment, the front-end interface enhances usability, accessibility, and analytical efficiency by offering:

(a)

User-friendly navigation—A digital map allows intuitive watershed selection and exploration of hydrological features without requiring specialized GIS software.

(b)

Real-time visualization—Drainage networks and rain gauge locations are displayed immediately, providing spatial context for decision-making.

(c)

Dynamic feedback—Contextual windows display watershed properties and preliminary outputs in real time, enabling rapid, iterative exploration.

(d)

Accessibility and platform independence—A web-based interface reduces reliance on local computational resources, broadening access for both experts and decision-makers.

(e)

Streamlined workflow—The interface provides a direct gateway to EBA4SUB and COSMO4SUB, ensuring a seamless transition from watershed selection to model execution.

Regarding the Back-End/Database, to construct the DEM, we selected the Copernicus EU-DEM at 25 m resolution. The standard pit-filling algorithm [43] was applied to eliminate topographic depressions, followed by a stream-burning procedure, in order to align the automatically extracted river network with the official digitized networks. Flow direction and flow accumulation were computed using the D8 algorithm [44].

Following validation of the DEM-derived drainage network against the digitized counterpart, the network was discretized by selecting coordinate points at 100-m intervals. These points function as active nodes within the interactive map interface and represent spatially explicit, user-selectable output locations.

The selected land use is based on the CORINE project (Copernicus Land Monitoring Service), year 2018. Starting from this land use, reclassification tables have been applied to link the generic land use to the corresponding numerical value of the Curve Number (CN) related to the methodology produced by the National Resources Conservation Service (NRCS), formerly the Soil Conservation Service (SCS), following what is described in detail in Grimaldi et al. [35].

Concerning the rainfall data and the DDF equations, observed rainfall data (daily and sub-daily resolution) were collected for all rain gauges located within the Lazio Region, in Central Italy. To estimate the DDF parameters, we employed the methodology developed under the VAPI (Valutazione delle Piene in Italia) project—an authoritative national-level initiative [45]. Conducted during the 1980s and 1990s by Linea 1 of the Gruppo Nazionale per la Difesa dalle Catastrofi Idrogeologiche (CNR-GNDCI), the VAPI project established a semi-institutional framework for regional-scale hydrometeorological hazard assessment across Italy. The method uses a two-component extreme value (TCEV) distribution for constructing DDF curves and applies regionalization based on climate-homogeneous zones, providing the statistical basis for infrastructure design and hazard assessment. Despite its age, the VAPI methodology remains the primary reference for the design of hydraulic infrastructure and for evaluating rainfall-induced hazards such as floods and landslides throughout the country.

Moreover, the database stored in the proposed platform provides all information related to the rain gauges—including their coordinates, time series length, basic statistical properties, and both observed and simulated time series using Virtual Rain.

The decision to pre-process and store this information offline serves two main objectives:

(a)

It significantly improves computational efficiency by performing resource-intensive tasks—such as pit-filling and stream-burning—in advance. This approach removes the need to integrate a full GIS engine within the platform and reducing the computational burden during user interaction.

(b)

It ensures consistency and spatial homogeneity of input data across the entire study area, supporting reliable, and reproducible model outputs.

Regarding the Computational Routines, the third environment consists of a group of ten modules developed in Phyton and C#. Hereinafter, their functionality is briefly summarized below:

• Module—Outlet Selection: This module allows the user to interactively select a specific point on the drainage network, which serves as the watershed outlet. The selection determines the upstream contributing area for hydrological modelling.
• Module—Watershed Delineation: Using the selected outlet and pre-processed flow direction and flow accumulation rasters, this module delineates the upstream watershed boundaries. The delineation is performed automatically, ensuring hydrological consistency with the digital terrain data.
• Module—Watershed Clipping: Once the watershed is delineated, this module extracts and clips all relevant spatial datasets (e.g., DEM, land use) to the watershed extent. This ensures that subsequent analyses are limited to the defined area of interest.
• Module—Watershed Attributes: This module calculates key physical and hydrological characteristics of the watershed, including area, elevations, centroid, concentration time (using the Giandotti formula). These attributes are essential for model parameterization.
• Module—Curve Number: Based on the land use distribution within the delineated watershed, this module computes the average Curve Number (CN) by applying NRCS [22] standard tables. The CN values are assigned on a pixel-by-pixel basis by linking land use classes—assuming a predefined Hydrological Soil Group—with the corresponding CN values specified in the official NRCS formulation.
• Module—DDF: This module retrieves or calculates Depth–Duration–Frequency (DDF) curves for the watershed area, using regional DDF parameters derived from the VAPI methodology (http://www.idrologia.polito.it/gndci/Vapi.htm accessed on 28 October 2025).
• Module—Rain Gauge Selection: It displays an interactive layer of available rain gauge locations, from which a representative station can be selected. Selection criteria may include spatial proximity or inclusion within the watershed.
• Module—Migration to EBA4SUB: This module packages and exports all necessary watershed data—geometry, CN, DDF curves, and outlet location—to the EBA4SUB environment. It supports detailed event-based hydrological modelling for runoff and peak discharge estimation, while also allowing users to modify key parameters such as CN and time of concentration.
• Module—Migration to Virtual Rain: It transfers user-selected parameters and watershed configuration to the Virtual Rain module. This module allows for the generation of stochastic or scenario-based rainfall time series to be used in further hydrological simulations.
• Module—Migration to COSMO4SUB: It prepares and exports input data for continuous modelling in the COSMO4SUB environment, including watershed geometry, rain gauge data, and climatic parameters. This module supports long-term hydrological simulations and water balance assessments.

The digital platform has been implemented across the Lazio Region in Italy (17,242 km²), with a specific focus on analyzing small to medium-sized watersheds—particularly those with a contributing area of less than 400 km².

To illustrate the platform’s operational workflow, we simulate a typical user interaction. The user may either begin with prior knowledge of a specific watershed (by inserting the outlet coordinates) or explore the region interactively via the map interface. By selecting a point along the drainage network, the user defines the watershed outlet. The system then automatically delineates the contributing basin, which is displayed on the map. Following delineation, a pop-up window (Figure 3a) presents key watershed attributes, including drainage area, concentration time, CN, and DDF parameters for four return periods commonly used in practical hydrology (50, 100, 200, and 500 years). Moreover, the platform allows users to view the application with the default parameters of the EBA4SUB model, using as input the information shown in the pop-up (Figure 3b). This output is intended as preliminary information for the user, providing a quick and easy characterization of the selected watershed. These outputs are generated during Phase 1 through the integration of back-end computational routines and pre-processed datasets stored in the offline database.

This initial phase is critical, as it streamlines conventional hydrological workflows and provides standardized inputs required by both the EBA4SUB and COSMO4SUB modelling environments.

In Phase 2, users pursuing event-based modelling can export the selected watershed configuration directly to the EBA4SUB online interface. All required inputs are automatically transferred and preloaded, allowing the user to proceed seamlessly with customized hydrological analyses (e.g., varying default parameters such as CN or concentration time).

For users interested in continuous simulation using COSMO4SUB (as described in Section 2), a long-term, sub-daily synthetic rainfall series is required instead of a conventional design hyetograph. By enabling the Rain Gauge option, all available monitoring stations are displayed. The user can then select a representative station—typically the nearest or most centrally located—which triggers the display of basic metadata such as geographic coordinates, observation period, and annual maximum rainfall values (Figure 4).

At this stage, the user is offered three available options:

• Execute COSMO4SUB directly using a pre-generated 15-min synthetic rainfall series retrieved from the offline database.
• Launch Virtual Rain module with observed daily rainfall data to generate new synthetic sub-daily time series.
• Launch Virtual Rain module with a 2000-year simulated daily rainfall time series to support advanced scenario modelling.

These pathways are supported by the platform’s comprehensive offline database, which integrates both observed and synthetically generated rainfall records. In Option 1, COSMO4SUB can be implemented immediately, as both topographic and rainfall inputs are pre-configured. Options 2 and 3 provide additional flexibility by allowing the user to generate or refine rainfall time series in Virtual Rain before proceeding with the COSMO4SUB simulation. Notably, at the end of the calculations, projects can be saved and subsequently loaded in a future session.

5. Conclusions

This manuscript presents a prototype digital platform specifically developed to support the estimation of design hydrographs in small and ungauged basins within the Lazio Region, Central Italy. The tool enables users to explore large geographic areas, delineate and characterize watersheds, and—if desired—seamlessly transfer input data to advanced hydrological models such as EBA4SUB (event-based model) and COSMO4SUB (continuous model).

At the core of the platform is the automation of traditionally complex and time-consuming pre-processing tasks—including watershed delineation, parameter estimation, and rainfall input preparation. Under conventional practices, these operations can require one to five full working days and advanced GIS expertise per watershed. The proposed system reduces this effort to just a few minutes, while enhancing reproducibility, consistency, and usability.

Key advantages of the platform include the following:

(a)

Drastic reduction in time and effort—Automated pre-processing of DEMs, Curve Numbers, time of concentration, and DDF parameters significantly accelerates workflows and lowers the technical barrier for practitioners.

(b)

Facilitated access to advanced modelling tools—The platform functions as an intuitive interface to sophisticated hydrological models such as EBA4SUB and COSMO4SUB, effectively lowering the entry barrier to research-grade tools. This enables practitioners to move beyond simplified empirical approaches and adopt scientifically rigorous methods for hydrological assessment and design.

(c)

Consistent and objective processing—Uniform execution of GIS operations (e.g., pit-filling, flow direction, stream burning) minimizes subjectivity and reduces the risk of user-induced variability in results.

(d)

Standardized regional input data—The integration of a centralized, pre-processed database (e.g., hydrologically conditioned DEMs, land use maps, regional DDF curves) ensures spatial and methodological consistency across multiple watersheds, supporting comparative and large-scale studies.

In summary, the proposed platform offers a scalable and operationally viable solution for bridging the gap between hydrological research and practice. By reducing technical complexity, accelerating model deployment, and standardizing input data, it encourages broader, more reliable, and informed use of hydrological modelling in engineering design, environmental planning, and risk management.

This work represents a pilot case study, with direct operational functionality currently limited to hydrological studies of basins within the Lazio Region, Central Italy. We envision that this study could serve as a reference model for numerous national and international institutions, allowing the platform to be extended both across Italy and to other countries. Such an expansion is straightforward, requiring only modest financial resources and collaboration with local stakeholders to define the necessary support and adaptations. Similarly, integrating other hydrological models available in different regions or for applications beyond small, ungauged basins could be achieved with minimal effort, further broadening the platform’s applicability and impact.