3 ms·

I learned this the hard way when engineering microwave links to withstand rain rates based on historical data. Many of my links that the calculators said would

I learned this the hard way when engineering microwave links to withstand rain rates based on historical data. Many of my links that the calculators said would have >99.999% availability saw a cumulative downtime in 2020 over 1hr due to much higher peak rain rates than historically measured. I now engineer the links to handle double the historically measured rain rates.

huhnmonster 6y ago
How does this manifest in costs? Is it noticeable? And does rain rate mean the frequency of distinct rain events over hours/days or the intensity of a single rainfall? Just curious, maybe you can shed some light on this
- sathackr 6y ago
  Sure -- It manifests in both cost and capability. For the same throughput you generally need more expensive equipment to obtain the same level of reliability. Where with the previous calculations it may cost X dollars for equipment, using the new calculations you may need to use a larger antenna or a more power transmitter, costing sometimes 2-3x, if it's even possible at all. Many times the link is just no longer viable at the required reliability. Rain rates are typically measured in mm/hr but it's an instantaneous rate, but what I've observed in 2020 is instantaneous peak rates easily 50% higher (indicated by signal attenuation levels) than 2019/2018. I'm not aware of any standard meteorological way to measure this as it seems most measurements are done as cumulative over minutes/hours/days, when the effects on microwave signals are in the seconds and minute scales. Certainly can't establish as causal relationship with climate change, especially since my observations are on the scale of just a few years, but it shows the danger in simply relying on low resolution historical 10/50/100yr data for infrastructure design.
kbutler 6y ago
Was that because the data was not actually peak _instantaneous_ rain rates, but rather rates averaged over a period of time (likely hours or longer)? Seems like microwave links would be most sensitive to instantaneous rate along specific paths, but any historical data would be measuring accumulated precipitation (averages over time) at specific locations.
- sathackr 6y ago
  This is true -- most vendors supply a calculator tool that attempts to take this into consideration. Over the past 10yrs or so I've found them to be generally accurate, but have noticed a significant deviation in late 2019-2020 from expected vs observed effects. I don't know what method is used but would expect it to be some sort of extrapolation from the average to predict peak rates

huhnmonster 6y ago

How does this manifest in costs? Is it noticeable? And does rain rate mean the frequency of distinct rain events over hours/days or the intensity of a single rainfall? Just curious, maybe you can shed some light on this

sathackr 6y ago
Sure -- It manifests in both cost and capability. For the same throughput you generally need more expensive equipment to obtain the same level of reliability. Where with the previous calculations it may cost X dollars for equipment, using the new calculations you may need to use a larger antenna or a more power transmitter, costing sometimes 2-3x, if it's even possible at all. Many times the link is just no longer viable at the required reliability. Rain rates are typically measured in mm/hr but it's an instantaneous rate, but what I've observed in 2020 is instantaneous peak rates easily 50% higher (indicated by signal attenuation levels) than 2019/2018. I'm not aware of any standard meteorological way to measure this as it seems most measurements are done as cumulative over minutes/hours/days, when the effects on microwave signals are in the seconds and minute scales. Certainly can't establish as causal relationship with climate change, especially since my observations are on the scale of just a few years, but it shows the danger in simply relying on low resolution historical 10/50/100yr data for infrastructure design.

kbutler 6y ago

Was that because the data was not actually peak _instantaneous_ rain rates, but rather rates averaged over a period of time (likely hours or longer)? Seems like microwave links would be most sensitive to instantaneous rate along specific paths, but any historical data would be measuring accumulated precipitation (averages over time) at specific locations.

sathackr 6y ago
This is true -- most vendors supply a calculator tool that attempts to take this into consideration. Over the past 10yrs or so I've found them to be generally accurate, but have noticed a significant deviation in late 2019-2020 from expected vs observed effects. I don't know what method is used but would expect it to be some sort of extrapolation from the average to predict peak rates