Tools4MSP: an open source software package to support Maritime Spatial Planning

MUC and CEA analysis

The Maritime Use Conflict (MUC) tool implements the COEXIST methodology to estimate the spatial distribution of the conflicts between sea uses. The inputs of the tool are: (i) the area of analysis; (ii) the grid cell resolution; (iii) layers of presence/absence for each human use present in the area (e.g., location of aquacultures, location of oil and gas platforms); (iv) an expert based characterization for each human use through four attributes (vertical scale, spatial domain, temporal domain and mobility). According to the attributes of each use three pre-defined rules, included in the COEXIST methodology are dynamically applied to estimate the potential conflict score for each pair of uses. The potential score varies from 0 (no conflict) to 6 (very high conflict). Afterwards, the area of analysis is subdivided into regular grid cells according to the specified resolution and, on each cell, information about spatial overlapping human uses are extracted. Finally, on each cell the total MUC score is calculated summarizing the potential conflict score for each pair of overlapping uses. The main output is a geospatial distribution of MUC score over the entire area of analysis. For a detailed explanation of the rules and algorithm we refer to Gramolini et al. (2010) and Barbanti et al. (2015).

The Cumulative Effects Assessment (CEA) tool implements the methodology described in Menegon et al. (2018a). Formally, we consider “CEA as a systematic procedure for identifying and evaluating the significance of effects from multiple pressures and/or activities on single or multiple receptors. CEA provides management options, by quantifying the overall expected effect caused by multiple pressures and by identifying critical pressures or pressure combinations and vulnerable receptors. The analysis of the causes (source of pressures), pathways, interactions and consequences of these effects on receptors is an essential and integral part of the process” (Judd, Backhaus & Goodsir, 2015).

The inputs of the Tools4MSP CEA tool are: (i) the area of analysis; (ii) the grid cell resolution; (iii) layers representing intensity or presence/absence of human uses (e.g., intensity of fishery and maritime transport, presence of aquacultures and oil & gas platforms); (iv) layers representing intensity or presence/absence of environmental components (e.g., seabed habitats, probability of presence of nursery habitats, probability of presence of marine mammals); (v) use-specific relative pressure weights and distances of pressure propagation; (vi) environmental component sensitivities related to specific pressures or more general ecological models that describe the response of the environmental components to a specific pressure. Similarly to the MUC tool, the area of analysis is subdivided into regular grid cells. Then, on each cell, information about the presence of human uses and environmental components are extracted. Afterwards, the geospatial layers on human uses are propagated and combined to estimate the geospatial distribution of different pressures (e.g., marine litter, underwater noise, abrasion) for the entire analysis area. Finally, the geospatial distribution of single and cumulative effects and impacts are estimated combining together the pressure layers and the environmental components layers through a sensitivity score.

The MSP-oriented integrated system

In Fig. 1 the interactions between the Tools4MSP Geoplatform, external application and potential end-users are presented. As a whole, they represent an integrated system of software, users and web services capable to effectively support MSP activities. Overall, the system can be described by four components:

The Tools4MSP Geoplatform

In Fig. 2 the detailed implementation architecture of the Tools4MSP Geoplatform is presented. The Geospatial CMS is the core of the system and is based on the GeoNode (GeoNode, Developers and Contributors Team, 2018) software, a Django-based web platform for developing community-based SDI. GeoNode facilitates the upload and management of geospatial datasets, making them discoverable and available via standard Open Geospatial Consortium (OGC) protocols and web mapping applications. GeoNode also allows users to automatically upload, describe and share the outputs produced by the Tools4MSP package.

The Tools4MSP package is Python-based open source software available on github (Menegon, 2018a). The Tools4MSP core library implements the algorithms for CEA and MUC geospatial modelling and the base functionalities to read the input geospatial datasets and configurations, to apply data transformation operations (such as normalization, aggregation, reclassification, gaussian convolution, geospatial filtering) and to manage and write model output results. Tools4MSP adopts a grid-based approach for efficient numerical computation of the geospatial models. The grid-based functionalities are provided through the general purpose RectifiedGrid library (Menegon, 2018b), that ensures direct integration of a multitude of different datasets and facilitates data preparation procedures. It simplifies the access and rasterization of multi-format geospatial data (environmental and anthropogenic datasets) and performs arithmetic and transformation operations on raster map layers. RectifiedGrid combines several FOSS Python projects: (1) NumPy and SciPy for efficient numerical computation (Van der Walt, Colbert & Varoquaux, 2011) and (2) rasterio and fiona (Mapbox, Rasterio Development Team, 2018; Fiona, Developers and Contributors team, 2018) for multi-format raster and vector data access through the Geospatial Data Abstraction Library (GDAL; Warmerdam, 2008). Interactive graphics for visualization of output results are based on the Bokeh visualization library (Bokeh Development Team, 2018).

Different users can have different modes of interaction with the Geoplatform: administrators perform the case study setup, by specifying the connectors to the geospatial repositories and defining the pre-processing chain for environmental and human use layers harmonization and utilization in the case study (Fig. 2, Case Study setup GUI).

A broad user community composed by decision-makers, planners, academics, research institutions, MSP stakeholders and the general public can use the case study configuration to run the build-in version of the Tools4MSP library implemented in the Geoplatform (Fig. 2, Webtool GUI). More in detail, the Tools4MSP Webtool GUI implements a four step workflow: (Step 0) Webtool selection (MUC or CEA); (Step 1) case study selection, choosing from a pre-set of case studies; (Step 2) case study configuration, optionally outlining a custom subregion of analysis or a custom combination of human uses, pressures, and environmental components; (Step 3) generation of geospatial and statistical outputs. The outputs are published in GeoNode and accessible through standard interoperable services. The produced reports, graphics and statistical outputs are archived in a dedicated data store catalogue and can be further visualized and re-used also by non technical stakeholders. In the results section an application of the Tools4MSP GeoNode plugin for case study setup and CEA analysis in the Adriatic Sea is provided.

Application of Tools4MSP GeoNode plugin

The Tools4MSP GeoNode plugin implements two sets of interfaces: the case study setup GUI and the Webtool GUI.

In Fig. 3 an example of GUI illustrating a case study setup is presented. The interface is intended for administrators and was designed to facilitate the configuration of case study input parameters. The GUI is organized in 4 sections: (i) basic parameters (Fig. 3A); (ii) human uses (Fig. 3B); (iii) environmental components (Fig. 3C); (iv) pressures (Fig. 3D). In the example the case study named “Adriatic Sea 2017” has a grid resolution of 500 m and the area of analysis is the Adriatic Sea (Fig. 3E). Figure 3F shows the expanded view of geospatial dataset set up for “Maritime Transport”. Specific data transformation operations have been configured through the “Pre-processing expression” field. The expression is written with a Python-based syntax that allows the user to select and combine one or more layers, apply filters, apply masking conditions, perform grid-cell-based arithmetic and other data transformations (e.g., normalization, logarithmic scaling, gaussian convolution). The resulting geospatial dataset is shown as thumbnail directly into the case study setup GUI (Fig. 3F).

The example reported in Fig. 3F refers to a human use (Maritime Transport) but is equally valid to environmental components or pressures.

The Webtool GUI is the standard interface allowing non-technical users (e.g., decision-makers, MSP stakeholders, planners) to perform CEA and MUC analyses starting from the pre-set case studies.

The Webtool GUI implements a four-steps workflow: (Step 0) Webtool selection; (Step 1) Case study area selection; (Step 2) Study area selection & Dataset configuration; (Step 3) Geospatial and statistical outputs. An example of the outputs (Step 3: Results) for the case study “Adriatic Sea 2017” is shown in Fig. 4. Through the GUI, the geospatial distribution of CEA for the analysis area as well as the statistical outputs can be explored and downloaded.

A complete example of MUC and CEA analysis through the Tools4MSP webtool GUI including a more in-depth investigation on its strengths and limitations in support of Maritime Spatial Planning is available in Menegon et al. (2018b).

Application of Tools4MSP package as stand-alone library

In this section we present a cumulative effects assessment based on the stand-alone Tools4MSP package applied for the Adriatic Sea. The case study set up has been downloaded from the Tools4MSP Geoplatform. It consists of a directory named “adriatic_sea” containing the input geospatial layers related to human uses, environmental components and pressures and environmental component sensitivities. An example of the case study set up is released within the Tools4MSP software package (https://github.com/CNR-ISMAR/tools4msp/tree/master/data/demo_case_study) and is available for test and demo purposes.

The case study is presented using a workflow implemented through the Jupyter computational environment including the following steps (Fig. 5):

• In [1] Import libraries: Tools4MSP CaseStudy class, rectifiedgrid using the rg alias, matplotlib and numpy. CaseStudy is a comprehensive class that provides the methods to setup a case study (e.g., specify datasets and input parameters) or alternatively to load a predefined case study, perform the analysis and manage the results.

• In [2] Load predefined case study for Adriatic Sea (1 km ×1 km resolution). After instantiating the CaseStudy class specifying the “adriatic_sea” case study, the methods load_layers and load_inputs are called in order to import the environmental, pressure and human use layers (including the layer metadata) and the input parameters respectively (e.g., sensitivity scores).

• In [3] Print example case study setup providing general information on spatial extent (in number of cells) and number of available layers (geospatial datasets). The input geospatial layers and the area of analysis layer (.grid parameter) are instances of the RectifiedGrid class. RectifiedGrid is the main class provided by the rectifiedgrid library and it is designed to represent and manipulate 2D georeferenced data arrays. RectifiedGrid provides the “.plotmap” method to plot the layer and other information (e.g., coastline, rivers) on a map.

• In [4] CEA analysis function. The CaseStudy class exposes the “.cumulative_impact” method to perform a Cumulative Effects Assessment (Menegon et al., 2018a).

• In [5] CEA geospatial results and graphical outputs for the Adriatic Sea. The map of the CEA result is visualized using the “.plotmap” method in combination with the distribution of CEA values of the grid cells.

Software details

In Table 1 the summary of the main characteristics and requirements of RectifiedGrid and Tools4MSP is presented.