GIS Dataset for Wetland-Class Trajectory and Hotspot Analysis in the Modeste Sub-Watershed, Alberta

Description

This dataset contains GIS layers and derived spatial outputs associated with a wetland-class trajectory analysis for the Modeste sub-watershed, Alberta, Canada, covering 2011–2024. The analysis was based on 30 m Agriculture and Agri-Food Canada (AAFC) Annual Crop Inventory (ACI) data, with annual land-cover observations reclassified into wetland and non-wetland states and assembled into pixel-level temporal trajectories. The dataset includes spatial outputs representing persistence-based wetland-class trajectories, including persistent wetland, persistent non-wetland, persistent gain, persistent loss, occasional and other dynamic trajectory classes. It also includes outputs from temporal and spatial statistical analyses using the Mann–Kendall test, Sen’s slope, and Getis–Ord Gi* hotspot analysis to identify spatially coherent candidate areas of wetland-class gain and loss. These data are intended to support reproducibility, visualization, regional environmental screening, and prioritization of locations for higher-resolution assessment or field verification. The mapped trajectories represent ACI-derived wetland-class patterns and should not be interpreted as independently confirmed ecological wetland gain or loss. The original AAFC Annual Crop Inventory datasets are not reproduced as original source products in this repository; users should obtain the source ACI data directly from Agriculture and Agri-Food Canada where required.

Steps to reproduce

1. Download the annual Agriculture and Agri-Food Canada (AAFC) Annual Crop Inventory rasters for 2011–2024 and the Modeste sub-watershed boundary. 2. Clip all annual ACI rasters to the watershed boundary, resample to 30 m using nearest-neighbor resampling, and align all rasters to a common snap raster. 3. Reclassify the ACI Wetland class (Value 80) as 1 and all remaining land-cover classes as 0. 4. Stack the 14 annual binary rasters and construct a 14-year binary trajectory for each pixel, ordered from 2011 to 2024. 5. Apply the supplied Python trajectory-classification script using a minimum persistence duration of 3 consecutive years. The script derives the persistence-filtered Genuine Sequence and assigns each pixel to one of the wetland-trajectory classes. 6. Run a pixel-wise Mann–Kendall trend analysis on the annual binary raster time series and extract Sen’s slope and p-value outputs. 7. Convert the Sen’s slope raster to points and perform Getis–Ord Gi* Optimized Hot Spot Analysis. 8. Filter the hotspot output using the Mann–Kendall significance criterion (p ≤ 0.05) to obtain the final statistically supported hotspot/coldspot layer. 9. Compare the resulting trajectory-class raster and significant hotspot layer with the corresponding files supplied in this repository.

Institutions

• University of Alberta

  Alberta, Edmonton

Categories

Licence