# Prompt guidance

Global Nature Watch is designed to help you quickly extract valuable, actionable insights from complex geospatial data sourced from Global Forest Watch (GFW) and Land & Carbon Lab. Think of it as an expert research assistant or colleague — the more precise and well-structured your questions (prompts), the more relevant and useful its answers will be. 

Here’s how to craft effective prompts to get the most value from the tool:

## 📝Core principles for effective prompting

### 1. Be clear and specific

Avoid vague or ambiguous language. Tell the tool exactly what information you are seeking. 

* Instead of: "Tell me about forests." 
* Try: "What was the total primary forest loss in the \[name of your defined area of interest] between 2018 and 2023, and what alerts were most frequent during that period?" 

### 2. Provide essential context

Always specify the geographic area you've defined (once selected or drawn in the tool) and the time period relevant to your query. You can also mention specific datasets if you have a preference. 

* "For \[your currently selected area] between \[start date] and \[end date], please summarize ..." 
* "Using land disturbance data, analyze the changes in ..." 

### 3. State your goal and desired output

Clearly indicate what kind of answer you expect. Do you want a summary, a comparison, a list, a trend analysis or an explanation? 

* "Summarize the key land cover changes in ..." 
* "Compare the rate of tree cover gain in region A versus region B." 
* "Identify the top 3 areas with the highest carbon emissions from deforestation in this region."

### 4. Iterate and refine

If the first answer isn't precisely what you need, don't hesitate to refine your prompt. Add more detail, narrow the scope or rephrase your question. It's an interactive process, and the AI learns from your refinement. 

* First prompt: "Show me deforestation in Brazil." 
* Refined prompt: "For the Brazilian Amazon, specifically within the Xingu Indigenous Park, how much deforestation occurred between 2024 and 2025?" 

### 5. Understand the platform's scope

Remember that the AI assistant processes and interprets the GFW and Land & Carbon Lab data within the platform. It does not have real-time external knowledge, personal opinions or the ability to conduct fieldwork. Avoid questions that require subjective judgment, future predictions or information not present in the integrated datasets. 

* Avoid: "Will deforestation increase next year in this area?" or "Why did the local government approve this specific land use change?" 

## 🤓 Prompting strategies and best practices

* <mark style="background-color:$primary;">**Start broad, then narrow down**</mark><mark style="background-color:$primary;">:</mark> If you're exploring a new area, begin with a general question and then ask follow-up questions to drill into specifics.&#x20;
  * "What are the major land cover changes detected in \[your area of interest] over the last 5 years?"&#x20;
  * Follow-up: "Can you provide more detail on the drivers of tree cover loss in that area, according to disturbance alerts?"&#x20;
* <mark style="background-color:$primary;">**Use relevant keywords**</mark><mark style="background-color:$primary;">:</mark> Incorporate terms specific to land monitoring and the datasets, such as: `tree cover loss`, `tree cover gain`, `carbon emissions`, `carbon sequestration`, `land disturbance`, `deforestation alerts`, `primary forest`, `grassland`.
* <mark style="background-color:$primary;">**Specify output format**</mark><mark style="background-color:$primary;">:</mark> If you prefer your answer in a specific structure, you can include this in your prompt.&#x20;
  * "List the primary causes of forest disturbance in \[area of interest] as bullet points."&#x20;
  * "Provide a brief summary paragraph of the carbon flux in \[area of interest]."&#x20;
* <mark style="background-color:$primary;">**Test and experiment**</mark><mark style="background-color:$primary;">:</mark> The best way to learn is by doing. Don't be afraid to experiment with different phrasings and question types.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://help.globalnaturewatch.org/resources/prompt-guidance.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.