📘 SECTION 2 — Investigation (10 marks)

Suggested length: ~400 words

✔ SEC–AlIGNED CHECKLIST

• Research on environmental context (drought, fire risk, soil moisture, canopy, etc.)
  
  

• Research on real monitoring systems (sensors, fire detection, drought systems)
  
  

• Research on embedded systems or related technology
  
  

• Research on models/simulations (weighted scoring, environmental prediction)
  
  

• Shows how research influenced design decisions
  
  

• Uses clear, credible, referenced sources
  
  

• Not too generic — shows depth and relevance

📘 SECTION 2 – Investigation (10 marks)

Suggested word count: ~400 words

📌 SECTION 2 – INVESTIGATION CHECKLIST (Aligned to SEC Marking Scheme)

The investigation must show:

✔ Environmental Research

• Drought
  
  

• Wildfire risk
  
  

• Soil moisture
  
  

• Temperature effects
  
  

• Canopy/light effects
  
  

✔ Research on Real-World Systems

• Fire-warning networks
  
  

• Sensor-based monitoring
  
  

• Environmental data collection
  
  

✔ Research on Embedded Systems

• Microcontrollers
  
  

• Data logging (CSV)
  
  

• Example projects relevant to the system
  
  

✔ Research on Modelling Approaches

• Weighted scoring
  
  

• Environmental modelling
  
  

• Why chosen approach is appropriate
  
  

✔ Connection Between Research & Decisions

• Sensor choice linked to research
  
  

• Data-logging choice justified
  
  

• Model choice justified
  
  

• Simulation choice justified
  
  

✔ Referencing

• All sources referenced
  
  

• AI tools acknowledged

🌟 MAX-MARK SAMPLE (≈400 words)

2. Investigation

2.1 Environmental Context and Problem Analysis

My investigation began by researching the environmental pressures identified in national and international forest-management documents. The Irish Forest Strategy (2023) highlights drought stress, declining soil moisture and rising temperatures as key factors contributing to increased wildfire probability. This closely matches the SEC project brief, which emphasises drought, canopy cover and fire risk as major environmental challenges facing forests.

This justified my decision to focus my project on drought conditions and wildfire risk rather than on less critical forest processes.

2.2 Scientific Factors Influencing Wildfire Risk

Research from the Climate Council of Australia explains that extended periods of high temperature combined with low humidity create optimal conditions for ignition and fire spread. These findings confirmed that environmental variables such as soil moisture, air temperature and sunlight exposure are among the most important indicators of wildfire danger.

This justified my choice to prioritise temperature, moisture and light level as the core variables measured by my embedded system.

2.3 Review of Existing Forest Monitoring Systems

I examined several real-world early-warning systems used to detect forest fires. Bosch’s IR/UV wildfire detection sensors use radiation patterns and temperature fluctuations to detect flames at long range. Although my project uses simpler sensors, this research demonstrated that modern fire-detection systems rely on analysing multiple environmental indicators rather than a single measurement. I also examined WWF’s forest-sensor networks, which monitor dryness and sunlight to detect drought stress and canopy loss.

This justified my decision to combine multiple sensor readings in my system instead of relying on just one type of input.

2.4 Embedded Systems and Data Collection Methods

To understand how environmental data is typically collected, I studied the Micro:bit Weather Station project (Micro:bit Educational Foundation, 2021). This project demonstrated how sensors can record temperature, light and moisture and store the data as CSV files for later analysis. This approach allows repeated and comparable data collection over time without needing a permanent computer connection.

This justified my decision to use the micro:bit’s built-in CSV data logging so that real sensor data could be used directly in my Python model.

2.5 Modelling and Prediction Techniques

I also investigated modelling approaches used in environmental science. Academic research on drought and fire prediction shows that when datasets are limited, weighted scoring models are often preferred because they are transparent, easy to interpret and scientifically defensible. This contrasts with machine-learning approaches, which require very large datasets to be reliable. Based on this research, I designed a weighted wildfire-risk model in Python where temperature contributes 50% of the risk score, moisture 30% and light 20%.

This justified my choice of a weighted linear model rather than a machine-learning approach, ensuring my predictions are reliable, explainable and suitable for the available data.

2.6 Impact of the Investigation on the Final Design

Overall, my investigation directly influenced all major design decisions in this project. It guided my selection of environmental variables, the use of CSV data logging, the choice of drought and wildfire risk as the simulated process, and the structure of my Python risk model. By grounding each decision in scientific and technological research, the final system closely reflects how real forest-monitoring and early-warning systems operate.

This ensured that the final artefact is scientifically informed, realistic, and strongly aligned with real-world environmental monitoring practices.

🌤️ MID-GRADE SAMPLE (≈200–250 words)

(less detailed, less justified, still correct but not exam-standard)

I researched forests and learned that drought and high temperatures increase the chance of wildfires. I found websites that said dry soil makes fires more likely, so I chose a moisture sensor. High temperatures are also connected to fires, so I used the Micro:bit temperature sensor.

I also looked at some fire-warning systems online and saw that they use sensors to measure the environment. This showed me that I needed to collect data regularly.

I researched the Micro:bit and looked at example projects. I found out that you can save data to a CSV file, so I used the same method in my project.

I also read about how to make models and found that risk scores are often used. So I created a simple model using the three values from my sensors.

From this research, I decided what sensors to use and how to make my model.

🔍 WHY THE MAX VERSION GETS 10/10 AND THE MID VERSION GETS 5–6/10 

🚀 SECTION 3 – PLAN & DESIGN (18 marks)

Suggested word count: ~600 words
What this section is really about:
The examiner wants to see your thinking BEFORE you build anything.
You must show:

1. You considered multiple options
   
   

2. You made informed, justified design choices
   
   

3. You understand stakeholders, user needs, technologies, and architecture
   
   

✅ SECTION 3 – What MUST Be Included (Checklist)

A. Design Objectives

Clear, specific goals of the artefact.

B. Options Considered + Justification

You must show at least TWO options and explain why you chose ONE.

Examples of options to compare:

• Microcontroller choices
  
  

• Sensors
  
  

• Outputs
  
  

• Data storage
  
  

• Model types
  
  

• What-if variables
  
  

• Feedback mechanisms
  
  

C. Stakeholders & End-User Needs

Who benefits and why?

D. Technologies Selected

Microcontrollers, sensors, Python libraries, embedded system components, HTML/CSS.

E. System Architecture Diagram

Your project’s “big picture” — inputs → processing → outputs.

F. High-Level Flow

What are the steps of your system from start to finish?

⭐ MAX-MARK EXAMPLE — SECTION 3 (Expanded)

(Students can use this as a structure, not to copy)

Design Objectives
My artefact aims to measure environmental variables (soil moisture, temperature, light), simulate drought cycles, model wildfire risk using Python, test what-if environmental scenarios, and provide an automatic adaptive response when risk becomes high. These objectives align with the core environmental themes in the SEC brief such as drought, wildfire risk, and biodiversity vulnerability.

Options Considered
I evaluated three microcontroller platforms: Arduino Uno, Raspberry Pi Pico, and the Micro:bit. Arduino offers excellent analogue input precision but requires an SD card module for CSV logging, adding cost and complexity. Raspberry Pi Pico has greater processing power but lacks built-in environmental sensors. The Micro:bit includes a temperature sensor, light estimation, built-in filesystem support, and a simple MicroPython interface, making it the best fit for rapid prototyping and classroom-based development.

For sensors, I considered a digital hygrometer but chose an analogue soil-moisture sensor because moisture is identified as the strongest predictor of wildfire ignition in Irish forest management studies (Department of Agriculture, 2023). I added a dedicated light sensor because the SEC brief specifically mentions light as a relevant piece of environmental data that “can help simulate real-world processes” such as canopy loss and heat buildup.

Stakeholders & User Needs
Stakeholders include forestry workers, farmers, national park personnel, and environmental agencies. They require early warning of environmental stress, wildfire conditions, and drought buildup. Their needs guided my decision to design a simple scoring model that is easy to interpret.

Technologies Used
I selected MicroPython for the embedded system due to its readability and Python compatibility. CSV logging was chosen because it is transparent, portable, and ideal for use with pandas in Python modelling. Python was chosen for the disaster risk model because it supports data manipulation and simulation through libraries such as pandas.

System Architecture

Sensors → Microcontroller → CSV logging → Python Model → What-if Simulation →

Adaptive Feedback (Buzzer)

High-Level Flow

1. User presses button to calibrate device
   
   

2. Sensors collect data every minute
   
   

3. Data logged to CSV
   
   

4. Python loads CSV, normalises values, assigns risk score
   
   

5. What-if scenarios adjust variables (temperature or moisture)
   
   

6. If risk > threshold, adaptive buzzer activates
   
   

⭐ MID-GRADE EXAMPLE — SECTION 3 (Expanded)

I planned to build a forest monitoring system using a Micro:bit. I chose it because it is easy to use and I have used it before. I planned to collect moisture and temperature and use them to calculate a fire risk score.

My idea was to store the data in a CSV file and then use a Python program to read it and calculate the score. I also planned to increase temperature or decrease moisture for my what-if scenarios.

I designed my system so that when the risk is too high the buzzer turns on.

🔍 WHY MAX MARK VS MID MARK?

Websites for making system architecture and flowchart

https://app.diagrams.net/ 

https://lucid.app/

https://www.canva.com/

FLOWCHART EXAMPLE, POSSIBLE TO USE SEPERATE FLOWCHARTS PER SECTION IE

ONE FOR THE MICROBIT

ONE FOR THE CSV CLEANING ON PYTHON

ONE FOR THE DISPLAY OF CURRENT CONDITIONS FROM THE CSV

ONE FOR THE WHAT IF SCENARIOS 

FLOWCHART SYMBOLS

🔍 WHY MAX MARK VS MID MARK?

🚀 SECTION 5 – EVALUATION (10 marks)

Suggested word count: ~300 words

✅ SECTION 5 – What MUST Be Included (Checklist)

✔ Strengths (specific)

✔ Limitations (technical, realistic)

✔ Future improvements (justified)

⭐ MAX-MARK EXAMPLE — SECTION 5 (Expanded)

Strengths
The system successfully collected consistent environmental data and stored it reliably in CSV format. My drought simulation reflected expected drying patterns. The wildfire-risk model behaved realistically, with risk increasing during hot, dry periods. The adaptive feedback mechanism worked without failure and demonstrated genuine environmental responsiveness.

Limitations
The moisture sensor occasionally fluctuated because analogue sensors are sensitive to electrical noise. My weighted model, while transparent, does not include humidity or wind, limiting realism. Light readings from the Micro:bit are approximate rather than calibrated lux measurements.

Improvements
I would add a humidity sensor and IR flame sensor to enhance prediction accuracy. I would also develop a more advanced risk model using multiple datasets. Adding Bluetooth to transmit alerts would make the system more practical for real forest monitoring.

⭐ MID-GRADE EXAMPLE — SECTION 5 (Expanded)

My system worked well and gave readings that looked correct. The buzzer turned on when the risk was high.

A limitation was that sometimes the moisture readings were not accurate.

If I continued the project I would add more sensors.

🔍 WHY MAX MARK VS MID MARK?

📚 REFERENCING GUIDE (REQUIRED BY SEC)

The SEC requires that ALL external sources — including websites, datasets, ideas, diagrams, AND AI tools — must be referenced.

M109 M110 26 Coursework

✔ Where to reference:

• At the end of the report AND
  
  

• Within the text where relevant
  
  

✔ Referencing Format (simple LC-appropriate)

Website

(Department of Agriculture, 2023, “Irish Forest Strategy”, www.gov.ie/… )

Technical Guide

(Micro:bit Educational Foundation, 2020, Weather Station Project)

Research Paper

(Climate Council, 2022, Fire Risk Research Article)

AI Tools (MANDATORY if used)

Include:

“ChatGPT (OpenAI, 2025) was used to assist with explanations and idea development. All final writing and decisions are my own.”

In-text referencing examples

Soil moisture is identified as one of the strongest predictors of wildfire ignition (Department of Agriculture, 2023).

The drought-cycle simulation approach is similar to examples used in Micro:bit weather-station projects (Micro:bit Foundation, 2020).