The most important instrument is observation. Before reaching for any technology, students should slow down and look. Below are the key things to notice for each event type. In all cases, note the location, the date, and the source of each observation.
High water marks on fences, walls, and trees. Debris lines showing how far the water reached. Changed stream banks and channel shape. Sediment deposits: where they are, how deep, what they contain. Damage to vegetation and infrastructure. Where the water came from and where it drained to.
The shape and size of the scar. The extent of the debris fan below it. Exposed layers of soil and rock. Changed drainage patterns around the failure. The vegetation on affected slopes compared with undisturbed slopes nearby.
What failed and what did not — and why. Fallen trees and what the root exposure reveals about soil depth and structure. Erosion at exposed sites. Changes to dunes, beaches, or coastal vegetation.
How far wave or surge evidence extends inland. Wrack lines, debris, and salt damage to vegetation. Sand and sediment movement. Changes to beach profile and dune structure.
Soil surface cracking and hardening. Vegetation dieback patterns and which species are affected first. Reduction in stream flow or wetland extent. Changed animal behaviour and movement.
Rainfall does not flood a river — the catchment does. How much water reaches a stream depends on rainfall intensity, how saturated the soil already is, how steep the land is, and what covers it. A cleared hillside responds very differently to the same rainfall as one under native bush. What does your local catchment tell you about the flooding you observed?
Slopes are held together by root networks, soil structure, and rock type. They fail when water adds weight and reduces friction faster than vegetation can absorb it. Steepness, underlying geology, and land use history all shape where slips happen. A slope that failed may show you its layers: topsoil, subsoil, parent rock. Read them.
Storm surge adds sea level on top of wave energy. The result can move sand, overtop dunes, and reach well inland of the normal beach. Where the surge evidence ends tells you the peak energy of the event at that point on the coast. Sediment moved by a storm does not all come back.
Drought is not simply the absence of rain. Soil moisture depends on rainfall, temperature, wind, vegetation cover, and soil type. Cracked, hydrophobic soil repels the first rain after a drought rather than absorbing it — which is why flooding can follow drought in the same season. What does your soil look like?
An extreme weather event is a single data point. Climate is the pattern across decades. A question worth asking after any event: is this becoming more common, more intense, or both? The answer requires long term data — which is why recording what you observe now matters beyond this classroom and this year.
These prompts are designed to be used after students have gathered their own evidence: observations, photographs, measurements, or testimony. The central task in each prompt is the same — compare what a gen AI chatbot says about how extreme weather works in general with what your class documented about how it worked in your specific place. The gap between those two accounts is where the learning lives.
Ask a gen AI chatbot to explain in simple terms why the flood, storm, or slip happened. Then share one thing you saw or heard from your community that the AI did not mention. What did the AI miss about your specific place?
Ask a gen AI chatbot whether events like this have happened before in New Zealand. Then ask an older member of your community the same question. Whose answer tells you more about your place, and why?
Ask a gen AI chatbot to explain how heavy rain turns into a flood. Draw a simple map or diagram showing the journey from rain cloud to the place you observed flooding. Does the AI explanation match what you saw?
Ask a gen AI chatbot what people do to prepare for and recover from this type of weather event. Compare the AI suggestions with what your community actually did. What did your community know that the AI did not?
Describe the event to a gen AI chatbot — location type, any rainfall or wind data you have, what you observed — and ask it to explain the physical processes. Where does the explanation match your field evidence? Where does it fall short?
Share your documented observations with a gen AI chatbot and ask what they indicate about the intensity and character of the event. Is the AI reading consistent with what you recorded? What would it need to know about your specific place to do better?
Ask a gen AI chatbot whether extreme weather events of this type are becoming more frequent or intense in your region of New Zealand. What evidence does it give? How would you test that claim using local sources such as NIWA data, community records, or long-term observations?
You have your own evidence and the AI has a general account of how this type of event works. Where do they match? Where does the AI general account fail to explain what you specifically observed? What does the gap tell you about the difference between general knowledge and local knowledge?
Describe your local landscape — land use, slope, vegetation, stream network, proximity to coast — and the specific event to a gen AI chatbot. Ask it to construct a causal account of why your place responded the way it did. Evaluate each step in the AI reasoning against your field evidence. Where is the account defensible? Where does it break down?
Your class has collected observations, photographs, measurements, testimony, and official data. Ask a gen AI chatbot to rank these evidence types by reliability for understanding a local extreme weather event, and to explain its reasoning. Challenge the ranking with what your combined evidence actually shows.
Ask a gen AI chatbot what land use changes, infrastructure responses, or community actions would most reduce your community's vulnerability to a repeat event. Evaluate each recommendation against your knowledge of the actual place. Which are feasible? Which assume conditions that do not apply here?
Your class holds direct observation of a specific event in a specific place at a specific time. Ask a gen AI chatbot what the limits of its knowledge are for understanding local extreme weather events. Then write a short reflection: what does your class's evidence contribute that the AI cannot replicate, and who in your community needs to know it?
| Level | Years 0–6 | Years 7–10 | Years 11–13 |
|---|---|---|---|
| 1 | Student names at least one thing that changed in the local environment as a result of the weather event, and can say whether they observed it directly or heard about it from someone else. | Student records specific observations about the event's visible effects, noting the source of each piece of evidence: direct observation, photograph, testimony, news report, or official data. | Student systematically documents the visible effects of the event across multiple evidence types, noting the location, time, and source of each observation, and identifies gaps in the evidence base. |
| 2 | Student explains in simple terms why the weather event caused the effects it did: "There was too much rain for the river to hold" or "The steep hill had nothing holding it together." The explanation connects cause to visible effect. | Student explains the physical processes behind the observed effects — rainfall intensity and catchment response, slope failure mechanisms, coastal dynamics, or drought indicators — using their specific observations as the evidence base. | Student constructs a causal account of the event connecting meteorological conditions to landscape response to community impact, using both field evidence and authoritative sources to support each step in the chain. |
| 3 | Student asks a gen AI chatbot to explain why the weather event happened and compares the answer with what they observed or heard from their community. Can name one thing the AI got right and one thing it missed about their specific place. | Student systematically compares a gen AI chatbot's explanation of the event with their own documented evidence. Identifies where the general account matches the specific observation and where it does not, and explains what the discrepancy reveals about each source. | Student analyses the gap between AI generated explanations of the event type and the specific evidence their class collected. Draws conclusions about what AI can and cannot contribute to understanding a local extreme weather event, and what that implies for emergency planning and community response. |
| 4 | Student explains what being in their place, or hearing directly from people who were affected, added that a video, news report, or AI explanation could not: the smell of floodwater, the sound of the wind, the sight of a familiar place changed, the words of someone they know describing what happened. | Student articulates what locally gathered evidence provides that secondary sources cannot: observations specific to this place, taken at this time, capturing the particular intensity and character of this event in this catchment, on this hillside, at this coast. | Student reflects on the value of locally gathered evidence in relation to modelled, reported, and AI generated accounts: what each contributes to understanding an extreme weather event, and why local knowledge, including student testimony, constitutes evidence that no external source can replicate. |
| 5 | Student names one thing their community did or could do to be better prepared for a similar event, and generates one question they would like to investigate about weather and their local place before the next event occurs. | Student formulates a monitoring or preparedness question based on their evidence: what they would measure, where, and how often, to know whether their place is becoming more or less vulnerable to this type of event over time. | Student designs an inquiry, monitoring plan, or community response proposal grounded in their field evidence: specifies the indicators, methods, and timescale, and identifies the form in which findings would be communicated to the people and organisations that need them. |