A Multi-Sensor, Multi-Location Dataset of Soil and Weather Conditions from Ireland
Md Babul Islam1,2,3
Declan Delaney 1
Antonio Guerrieri2,3
XXXXXXXXX 3
Others 4
1 University College Dublin, Ireland
2 University of Calabria, Italy
3 ICAR CNR, Italy




Overview: Our data includes a diverse range of scenarios, both indoor and outdoor features, featuring various locations and various depths in the soil, eg, 5, 10, 20 depth in the soil. The outdoor features such as temperature, humidity, rain, etc. The first row shows the time, and all other features names such as VWC, EC, and Temp. and others features while the second row demonstrates the features' values.



Paper

Soil Dataset

Code (Coming soon)


Dataset Abstract

We introduce a new dataset, Soil Moisture data, and weather data, comprising 10 locations/scenarios captured from various depths and weather views. Our dataset has diverse patterns and challenging variations, such as different locations and weather conditions, providing a robust foundation for training superior models. We conduct an in-depth analysis of recent representative models using our dataset and highlight its potential in addressing the challenges of soil moisture forecasting across diverse and evolving scenarios in Ireland.


Dataset Overview

Sensors

Soil, Weather

Location

10 Loactions in Ireland

Weather Sensors

Temperature, Rain, Air, Humidity, Pressure_Avg

Soil Sensors

VWC, EC, and Temperature

Scenarios

Real

Time Duration

∼4 Years More

Organization

UCD, Met Éireann

Duration

30 Minutes



Click To view extended details of our soil sensors dataset. Our dataset has a broad range of features, including location and weather.
Sesnors Features Deatil about the features Measure Unit
Weather Sensors 🌤️Weather Sesnors Humidity Sensors Numeric
Temperature Celsius
Air Tempeatures Celsius
Rainfall Event mm
Grass or Surface Temperature Celsius
Atmospheric Pressure hPa
Locations Tulamore --
Lynos Farm --
Solohead --
Johnstown Castle --
Dowth --
Dooary --
Curtins Farm --
Cregduff --
Ballycanew --
Athenry --
🌱Soil Sensors Features Volumetric Water Content (VWC) m³/m³
Electrical Conductivity (EC) dS/m
Soil Temperature Celsius
Depths 5cm --
10cm --
20cm --
30cm --
50cm --
100cm --

Click to view some graphical statistics results from Our Soil Moisture Dataset.



The statistics of VWC dataset include: (left) a breakdown of mainly various soil locations and their respective weather, (middle) a boxplot illustrating frame number variations across scenarios in VWC training set, and (right) the distributions of train/test split across scenarios.


Dataset Evaluation

Based on the experimental results, we can deduce that a model trained on intricate real-world scenarios exhibits superior generalization. This stems from the fact that real-world models are frequently influenced by the surrounding environment, encompassing elements like fluctuating traffic patterns, dynamic electronic displays, and the movement of trees in the wind. The model must discern the nuances of anomaly detection within a dynamic environment and comprehend the dynamics of objects and/or performing subjects within it. VWC dataset provides a comprehensive representation of real-world scenarios.

TABLE I — SoilFormer Result Comparison with Non-Pruned and LSTM Models: 5cm Depth, Lyons Farm
Model Dataset MAE RMSE R2
LSTMTrain.........
Validation.........
Test.........
SoilFormer (Non-Pruned)Train.........
Validation.........
Test.........
SoilFormer (Pruned)Train.........
Validation.........
Test.........

Cross-scenario evaluation

Train Test Micro Macro
ShT UCSD Ped2 57.38 58.36
ShT CUHK Avenue 69.98 78.32
ShT MSAD 63.92 64.92
MSAD UCSD Ped2 70.35 65.74
MSAD CUHK Avenue 79.57 84.49
MSAD MSAD 69.96 69.60
Evaluations on cross-scenario setups. We use FSAD and SA2D (ours) for training on ShanghaiTech (ShT) and MSAD, respectively.


Download

Please read the terms and conditions here and complete the below linked request form. Upon review, you will be notified of the outcome of your request via the provided email.

v1.0



Soil Dataset in Ireland


Paper


Soil Former: A Customized Transformer for Soil Moisture Forecasting in Irish Scenarios


Cite

@misc{babul024soilsensorsvwcforecasting,
      title={A Soil Moisture Forecasting: A Concise Study and a New Dataset}, 
      author={Md Babul Islam and Declan Delaney, Antonio Guerrieri, and others},
      year={2024},
      eprint={2402.04857},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/}, 
}


Contact

Please contact the following people for any inquiries related to the dataset or the paper.