Forewarned is forearmed

Whatever their domain – land, asset, production or risk – managers are hungry for good intel about the future. That’s where NIWA’s customised forecasting comes in, writes Colin Barkus.

Hawke’s Bay’s rural firefighters won’t forget the summer of 2009 in a hurry. Throughout January, desiccating northwesterly winds pushed the fire risk a little higher each day. Then, on the morning of 1 February, relative humidity plunged below 30 per cent and the temperature soared towards 30°C, arming an incendiary time bomb in the space of a few hours.

Sure enough, by late afternoon, quiet coastal settlements near Mahia Peninsula were surrounded by blazing scrub. No-one is certain of the cause, but crews soon realised they’d turned out to no ordinary wildfire. By the time the last hot spots were hosed into submission a week later, 110 residents had been displaced, 240ha of land had been charred and eight buildings – mercifully, only one of them residential – had been razed. The inferno cost around half a million dollars to quell.

“Without a doubt, we were stretched during that first day at Mahia,” says Gary Lockyer, Manager of Rural Fire Operations for the National Rural Fire Authority (NRFA). “That fire was out of control during the first six hours. A community faced significant threat.”

National Rural Fire Authority 

Part of Lockyer’s job is to calculate risk using the New Zealand Fire Danger Rating System, a model that generates fire-risk and behaviour indices. The country’s 73 regional rural fire authorities rely on that information to marshal firefighting resources, inform the public, manage restrictions and bans and issue fire permits.

It’s an awkward juggling act, but it’ll get considerably easier this July, when EcoConnect-Fire goes live. NIWA’s new online tool will give fire managers earlier warning when environmental conditions point to another Mahia, keeping Lockyer and his regional managers posted with daily (later to become hourly) local fire-weather index forecasts up to six days ahead. They’ll be backed by 2-, 6- and 15-day weather forecasts, and a range of longer-term climate analyses of the fire season ahead.

Back in 2009, NRFA’s risk management relied on software adapted from a Canadian rating system developed in the 1960s. Combining data from around 100 NRFA-operated weather stations, the software generated a daily fire-weather status report, then despatched it to regional managers each morning. That report – combined with their assessment of local conditions – underpinned each manager’s operational decisions.

“While it was pretty clear the Bay was getting dry,” recalls Lockyer, “the system wasn’t capable of warning us, ahead of time, that widespread extreme conditions would develop so rapidly that day.” With more warning, fire chiefs would have seen the need for backup.

Instead, he says, “we were playing catch-up from the outset".

Data to the rescue

Lockyer and his regional teams tread a fine line: “The public want to be able to go about their business; to visit recreational areas … use machinery and chainsaws in rural areas and forests. We can’t just slap prohibitions on, (because) they affect the community. But when conditions pass a trigger point, we need to apply restrictions and deploy resources proactively and confidently.

“So we need solid science to support whatever decisions we make. That’s where EcoConnect-Fire will come into its own.”

NIWA atmospheric scientist Dr Michael Uddstrom has led the development of EcoConnect services since its launch in 2007. He says EcoConnect “bridges the gap between NIWA’s science and businesses – like NRFA – that may be affected by certain environmental conditions.

Find out more about EcoConnect

“Once we understood the NRFA’s requirements, we sat down with our research partner, Scion, to design a system that would overcome the existing limitations.”

Scion website 

Uddstrom and Lockyer knew the NRFA’s weather-monitoring network needed work, so they incorporated some 80 key NIWA climate stations and upgraded a number of NRFA sites as well. “That significantly increased the density of highquality data available to produce more accurate forecasts in areas of greatest concern,” Uddstrom says (see sidebar: 'An electronic vigil').

NIWA’s advanced numerical modelling capability does the rest (see sidebar: 'NIWA forecasts'). EcoConnect-Fire feeds data from the monitoring network into a sophisticated weather prediction model that generates forecasts for temperature, relative humidity, precipitation, wind speed and wind direction across a high-resolution grid of the entire country. These will inform the NRFA’s new Fire Weather System, which will issue fire-weather indices and weather forecasts – along with climate analyses – to regional fire managers.

Each forecast, and all observational data, will be stored, allowing past emergencies – and resulting operational decisions – to be reviewed. The tool will also display realtime weather data where available, giving managers an instant snapshot of current conditions in their area. Automatic SMS or email alerts will signal when conditions are forecast to pass user-defined thresholds.

“Every decision we make and resource we deploy during the fire season involves time and money,” says Lockyer. “EcoConnect-Fire will allow us to act with justifiable confidence – both in advance, and when we’re in the midst of a fire.

“Even if an extreme situation doesn’t eventuate as forecast, we’ll know we’ve acted with the backing of the best science available.”

[article continues below video]


Weather forecasting - Let's make it better! from NIWA on Vimeo.


Bacterial warfare

It’s weather of a very different kind that concerns Charlotte Hardy.

When a prolonged spell of damp and windy conditions sets in, an insidious little bacterium called Pseudomonas syringae pv.actinidiae (better known as Psa) gets moving. The virulent form of Psa – Psa-V – causes canker in kiwifruit, manifesting as leaf spot and a viscid, orangey-red discharge from the trunk or leaders of the vine. At worst, the canker can kill the vine or render it susceptible to attack from pests and other diseases.

Hardy works for Kiwifruit Vine Health (KVH) Inc., the agency charged with minimising the impact of Psa-V on the New Zealand kiwifruit industry, and helping affected growers rebuild their orchards. As Technical Research Coordinator, her focus is on the science that will help halt Psa-V’s destructive advance.

Psa-V was the first major disease “that’s really threatened the industry,” says Hardy. When it was positively identified in Te Puke in 2010, “the industry went into panic. Many growers felt angry and let down".

With estimates of potential losses reckoned at some half a billion dollars, the industry galvanised: KVH, the New Zealand Institute for Plant & Food Research Ltd and Zespri teamed up to try to curb the blight. Plant & Food scientists, who’d been running a kiwifruit breeding programme for many years, set to work modelling risk and identifying Psa-V’s favoured environmental conditions.

“A shower here or there is not the real concern,” Hardy explains. “It’s when you get leaf dampness lasting 48 hours or more that the bacteria become highly mobile. They have two flagella – little wiggly tails – so they can swim on the surface of wet plants, and enter the vascular system through small openings, like stomata on the leaves and lenticels in the stems.

“Wind is also a factor. The bacteria can get airborne, and blow from vine to vine with rain splash. So wind and prolonged dampness – they’re the conditions we’re looking out for.”

Uddstrom knew EcoConnect could help.

“We’d worked with Zespri to produce a wind plume model that estimated how Psa-V may have spread since the first positive identification in Te Puke,” he says. “Once Plant & Food had developed the risk model, we knew we could forecast that risk through EcoConnect.”

But first, weather-monitoring capability had to be stepped up: four new stations were installed for KVH, then NIWA added five more in kiwifruit country and upgraded another four.

They now feed data to EcoConnect, to generate weather forecasts out to two days ahead at high resolution, and six days at less-accurate lower resolution. Critical locations get 15-day weather outlooks. The Psa-V risk model issues a simple six-level risk index, with 1 indicating minimal risk of an infection spreading, and 6 indicating extreme risk. Growers can access their local risk index and weather forecasts at any time on KVH’s website.

“We encourage growers to undertake orchard management activities during the low risk periods,” explains Hardy. “We don’t want them working in the rain, when water on tools can spread bacteria between vines.

“And when bad weather is in the forecast, we encourage them to spray in advance for protection.”

She says the service is proving increasingly popular, but there’s more to learn: “Recent indications are that heavy frost may play a role in spreading Psa-V, by breaking plant tissue. That’s not a variable in the model at the moment, but as EcoConnect can already forecast frost, the risk model could be improved as we understand more about the disease and its epidemiology.”

That sort of feedback is easily incorporated, says Uddstrom: “We call EcoConnect a ‘bridge’ because a two-way flow of information is essential. EcoConnect products will evolve as NIWA’s science advances, but also as users present us with new challenges and new information.”

EcoConnect’s forecasting capabilities go beyond weather and climate. Models predicting river flow, sea state, tides and storm-surge are incorporated, offering customised multi-hazard forecasts to anyone who might have to base management decisions on environmental conditions.

It was even used to forecast lake conditions during the World Rowing Championships held at Lake Karapiro in 2010 – and organisers praised it as the most accurate forecasting they had available anywhere in the world.

Find out more about our work forecasting lake conditions at the World Rowing Championships in 2010

For Gary Lockyer and Charlotte Hardy, the stakes are high. They might have different ideas of what a bad forecast looks like, but they agree that environmental forecasting is critical to managing risks and seizing opportunities.


Sidebar: NIWA forecasts

NIWA's forecasting system EcoConnect is powered by sophisticated numerical forecasting models. Data from NIWA's environmental monitoring network, and from other sources including satellites, weather balloons, ships and aircraft, feed a series of complex computer programmes processed by NIWA's IBM p575 Power6 supercomputer in Wellington. The programmes use elaborate mathematical formulae to mimic natural systems and processes, generating discrete forecasts for locations as close as a few kilometres apart. The resolution is improving all the time. Individual models predict atmospheric variables; land surface attributes such as soil moisture; ocean wave height, direction and period; tides; storm-surge and river flow. Outputs from different models can be combined to produce tailored multi-hazard forecasts. All data fed in from the monitoring network are quality controlled, and the model forecasts are continuously validated against actual data. Where real-time data are available from weather stations, EcoConnect users can compare actual values within a range of parameters, versus those forecast, just by looking at their computer screen.

Sidebar: An electronic vigil

NIWA runs a network of about 330 solar-powered environmental monitoring stations around the country, silently taking the environmental pulse every few seconds – day and night. Those measurements are sent by satellite or cellular phone network to the New Zealand Climate Database held in Wellington, or direct to end users via the Internet. Electronic sensors typically monitor temperature, rainfall, relative humidity, wind speed, wind direction and barometric pressure, but selected stations also log sunshine hours, solar radiation, grass temperature, soil temperature, soil moisture, leaf wetness and snow density. As well as receiving real-time data, the New Zealand Climate Database holds historical information dating back to the very first climate observations made in the 1850s – a trove of information used by NIWA and many other national and international organisations. It also feeds into NIWA's numerical forecasting models.


Environmental Forecasting Scientist