The Gulf of Riga DGA Workflow

[!NOTE] 📹 Video Reference 21:29 Detailed Review of the Workflow Components

This chapter details the specific “Data Gathering and Analysis” (DGA) workflow used in the case study. The workflow integrates eight distinct tools into a pipeline.

Input Data

The workflow starts with two datasets:

graph LR
    A[Input Data] --> B[Pre-Processing]
    B --> C[Spatial Aggregation]
    B --> D[Temporal Aggregation]
    C --> E[Trend Analysis]
    D --> E
    E --> F[Results]

Point Data: In-situ measurements of Secchi depth (water transparency).
Polygon Data: Assessment unit polygons defining the spatial boundaries of the regions being studied.

Pre-Processing

Spatial Aggregation

Goal: Assign scattered data points to specific administrative or geographical units.
Process: Points are mapped to “Assessment Unit” polygons. A new attribute is added to each data point indicating which unit it falls into.

Temporal Aggregation

Goal: Account for seasonality (crucial in the Gulf of Riga).
Process: Based on the visit date, each point is assigned a Season (e.g., Spring, Summer).

Data Analysis Steps

Calculate Mean Values (Group 1)

Calculates the average transparency for granular groups defined by:

Season
Year
Assessment Unit ID
Location

Calculate Mean Values (Group 2)

Further aggregates the data to get a single value per unit-year-season.

Output: A structured CSV file where each record represents a transparency value for a specific unit, year, and season.

Time Series Interpolation

To prepare for trend analysis, the data must be continuous.

Exclusion Criteria:

Units with < 10 data points are excluded.

Units where > 80% of data is missing (compared to the time series length) are excluded.

Interpolation: For the remaining units, missing values are statistically interpolated to fill gaps.

Mann-Kendall Trend Analysis

The final analytical step.

Method: A Mann-Kendall test (non-parametric statistical test) is applied.
Purpose: To detect if there is a monotonic upward or downward trend in the Secchi depth variable over time.

Visualization

The workflow automatically generates:

Interactive Map: Showing the assessment units.
Bar Chart: Displaying Kendall’s Tau values. A negative Tau indicates a decreasing trend (darkening), while positive indicates clearing.