The State of Weather Forecasting Technology in Pakistan

Extreme weather-related events have resulted in over 2 million deaths and $4.3 trillion in economic losses between 1970 and 2021 according to World Meteorological Organization. But despite the deepening economic losses, early warning systems have saved lives as recorded deaths in 2020 and 2021 are found to be lower than the previous decade’s average.

World Bank projects Pakistan’s Gross Domestic Product to decline by at least 18 percent to 20 percent by 2050 due to climate change and environmental degradation. Its induced reductions in wheat and rice crop productivity are to cost $19.5 billion in real GDP by mid of the century.

Despite emitting less than 0.9 percent of global emissions, Pakistan is the 8th most vulnerable country to climate change and incurred more than $30 billion in losses last year alone.

Amid this looming threat of climate change, it’s critical to shift the conversations beyond mere disasters, their effects and rising temperatures. Weather forecasting is the frontline defense against the unpredictability of nature and in addition to being a matter of life and death for thousands, nearly all business sectors and average individuals require accurate weather predictions for their day-to-day decision-making.

Where Does Pakistan Stand?

Our biggest challenge is the lack of data collection points because the more data we will have the better will be the predictions and the resulting future modelling.

The satellite-based data that is largely available is not as accurate for precipitation and Pakistan has no more than 110 weather stations under the Pakistan Meteorological Department for perspective, Japan which is nearly half of the size of Pakistan has approximately 6500 weather stations.

Pakistan has reportedly only two Doppler weather radars, but Pakistan Meteorological Department is always somewhat reluctant to share the data with business entities.

That lack of data points is the primary reason that we often hear very vaguely phrased weather advisories like a ‘possibility’ of rain over several hours of time frame, something that can be considered accurate regardless of the outcome, although Met Department claims its short-term forecast to be accurate up to 90 percent.

“Pakistan has vast kind of topographies from the heights of K2 to deserts at sea level and from plains and plateaus to a lengthy coastline and that creates challenges for weather forecasting with limited data points” commented Tabbish Mahmood, Head of Digital Services at Syngenta and Co-founder at Bakhabar Kissan.

Bakhabar Kissan is Pakistan’s one of the first and biggest Agritech companies with at least 20 million farmers’ user base and out of their 3 million monthly engagements, 70 percent are for Weather advisories.

“Farmer is ready to pay for these services, but logically they should not be charged as these are necessities though there can be premium offerings around it” added Mahmood.

He also said that weather is all about historical data, now cast and the forecast and Pakistan will eventually need thousands of weather stations for more accuracy and better coverage of 0.8 million sq kilometres. He also recalled frost damage to potato crops this winter and last year’s floods and said that there are lot more needs to be done by everyone to fight the rising threat of climate change to our food security.

The Value Chain of Weather Forecasting

Weather forecasting is a vast subject and goes way beyond just displaying temperatures and rain possibilities, and it’s easier to understand if we divide it into a value chain just like any other industry. Another reason that it is critical to understand its various verticals is that no single entity public or private can solve and capture all of it, so we need a lot of specialized players and collaboration.

The foremost, as mentioned earlier, is the data collection points, and it can be done through both weather stations & radars and satellite observations, but ground-based observations are considered superior in certain aspects over satellite-based information.

There are dozens of weather parameters that these data points collect like temperature, pressure, humidity, wind speed and direction, precipitation, cloud cover, visibility and even lightning which is often regarded as one of the most accurate indicators of monitoring and predicting thunderstorms and can assist in issuing warnings.

Then comes the weather data processing and analysis where it’s integrated with the historical data to initialize forecasts. The forecasts can be either short-term (few hours or days), medium range ranging from a week to weeks over a region or a continent or long-term modelling that helps in risk management. This data is later visualized in understandable maps and graphics for dissemination and expert analysis.

Dissemination of this information is also critical through all available mediums including broadcast media, websites and apps. The advisories are also modified for respective industries and use cases like aviation, agriculture, transportation or individual lifestyle choices. Coming back to Pakistan, we have traditionally been behind in nearly all areas, but things may very well start to change!

Weather Walay

Junaid Yamin left his job at the Australian Department of Health and returned to Pakistan in 2021 to establish Pakistan’s first private digital weather service provider that has built at least one hundred of its own weather stations to provide hyper-local weather forecasts. Their app is centred around lifestyle users, but they collaborate with Bakhabar Kissan and other entities as well for specialized advisories.

“When I looked at the data and how badly weather is impacting Pakistan, it was one of the biggest motivations to build something and centre it more around actions than words alone,” said Yamin, Co-founder and CEO at Weather Walay.

He said that they have mapped 620 weather zones in Pakistan in consultation with the veteran Climate Scientist Dr Mohammad Hanif and four decades of data and topographical information which is the bare minimum to cover the whole country.

He pointed out that their aim is to roll out one thousand weather stations to fix the lack of data points and build a central repository of multiple kinds of climate data over which long-term climate modelling can be done through artificial intelligence and machine learning. Weather Walay also won the GSMA Innovation fund and has signed MoU and is collaborating with multiple universities and companies including Jazz, Ufone, Bayer and Valley Irrigation.

  Team Weather Walay

It collects eight primary parameters temperature, humidity, wind speed, wind direction, pressure, rain accumulation, UV Index and Solar radiation, but the derived parameters like feels like temperature and rain rate (rain accumulation over time) are up to sixty. For example, if the rain rate tells you that 100 mm of rain has occurred in the last hour that means it’s a cloudburst and will be followed by flooding.

“We are using multiple international numerical weather models in our solution, and bring the best and most accurate information to the end user. Each model’s accuracy varies based on seasons and locations. In other words, our impact-based forecasts in our app automatically adjust to the most accurate model suitable for your particular location”, added Yamin.

He revealed that they are also educating the public that the correct metric to look for in summer is ‘Temperature Feels Like’ and ‘Wind Chill’ in winter which merges temperature and wind speed.

When asked about future business plans in light of low consumer purchasing power in Pakistan, he said that nearly all industries require accurate weather forecasts and around eight companies in Agriculture, Energy, Insurance and related sectors are in talks with them to have their APIs, so there is no shortage of demand.

Weather Walay already has six climate-related projects, one of which aims to turn the Cholistan desert green and will start next month. But Yamin said that they are strategically optimizing their pace to manage the resources and carefully looking to expand their team of seventy people.

Future Ahead

It is going to take a lot more specialized players to address the need of 240 million people and dozens of economic sectors over 0.8 million sq kilometers and more than that, it will need collaboration as it will be logically more efficient for multiple weather companies to cover the area and share the data to grow together.

Building early warning weather systems, dissemination of advisories to masses, import assistance for hardware, the establishment of more radar-based weather stations and better allocation of climate financing will require round-the-clock efforts by the government and private sector because the threats of climate change are growing with every passing hour. We are not prepared for the least of it.