Traditional leak detection and repair relies on someone walking a site every quarter or month with an optical gas imaging camera or a Method 21 sniffer, checking one component at a time. A continuous monitoring alternative flips that model on its head: instead of a periodic human survey, an always-on network of fixed sensors watches the whole site continuously and raises an alert when emissions rise above what normal operation should produce. This guide explains what these programs are, how the technologies fit together, and why turning a wall of raw sensor data into a defensible leak alert is fundamentally a data-validation and reconciliation problem rather than a hardware one.
Continuous Monitoring Alternative in one line: A continuous monitoring alternative to LDAR uses fixed or remote sensors - point detectors, open-path lasers, and aerial or satellite screening - to watch a whole site for methane emissions all the time, replacing the periodic component-by-component walk-through of optical gas imaging or Method 21. Regulators can approve it as an alternative test method when the operator proves the monitoring, alert logic, and follow-up investigation reliably find and fix leaks at least as well as the survey it replaces.
Conventional LDAR is a snapshot approach. A technician visits on a fixed schedule and inspects each valve, flange, connector, and seal, so a leak that starts the day after a survey can run for weeks or months before the next visit catches it. The method is thorough at the component level but blind between visits, and it scales poorly across a fleet of dispersed sites because every survey is a truck roll and a person-day. Continuous monitoring exists to close that time gap: rather than finding leaks at inspection intervals, it aims to detect an emission shortly after it begins.
The trade-off is a change in what is being measured. A Method 21 sniffer or an optical gas imaging camera localizes a leak to a specific component, which is exactly what a repair crew needs. A fixed continuous monitor usually measures concentration or a site-level emission rate at some distance from the source, so it detects that something is leaking and roughly where, but not always which flange. That is why continuous programs pair broad, always-on detection with a targeted follow-up survey once an alert fires - the sensor network tells you a site needs attention, and a technician then pinpoints the component. The two approaches are complementary rather than identical.
Because the whole point is to find abnormal emissions, a continuous system has to know what normal looks like. Wind carries a plume differently every hour, tanks breathe as temperature changes, and a compressor blowdown is a real but permitted release. A useful monitoring program therefore does not just measure gas; it compares the measurement against expected behavior for the current conditions and only escalates when the reading is genuinely anomalous. That comparison, not the sensor itself, is what separates a nuisance alarm from an actionable leak alert.
No single technology covers every leak size and every location, so continuous programs stack several. Point sensors sit at fixed positions around a facility and measure methane concentration in the air passing them; they are inexpensive enough to deploy in numbers and respond quickly, but each one only sees gas that the wind happens to blow past it. Open-path instruments shoot a laser or infrared beam across a span and measure the total gas along that line, giving broader coverage from fewer devices and catching plumes that would slip between point sensors.
Above the site sits the screening layer. Aerial surveys by aircraft or drone and satellite passes cover large areas and are strong at catching large releases - the super-emitters that dominate total emissions - but they see a site only intermittently and have a detection floor below which smaller leaks go unseen. In a well-designed program these layers hand off to each other: satellite or aerial screening flags a large anomaly across a region, fixed ground sensors provide the continuous local watch, and a human survey confirms and locates the source for repair. Each layer compensates for the others' blind spots.
Turning that layered data into a decision requires more than a threshold on a single number. Wind speed and direction determine whether a downwind sensor could even see a given source, so meteorological data has to be fused with the gas readings to estimate an emission rate and back-calculate a likely location. Multiple sensors seeing a coordinated rise, consistent with the wind, is far stronger evidence than one sensor spiking alone. This inversion - going from scattered concentration readings plus weather to a credible source and rate - is the analytical heart of a continuous alternative, and it is where most of the engineering effort actually goes.
A continuous monitoring alternative generates a relentless stream of data: many sensors, sampling frequently, across many sites, forever. Raw counts of that data are useless to a compliance manager. What is needed is a validated, time-aligned record that can be trusted, alert logic that fires on real anomalies without drowning operators in false positives, and an auditable trail showing that each alert was investigated and resolved. Those are the same capabilities a cloud SCADA and monitoring platform already provides for pressures, flows, and temperatures - which is why methane monitoring increasingly lives in the same supervisory layer as the rest of the site's data.
Data validation comes first. A sensor that has drifted, frozen, lost communication, or reported bad quality must be caught and excluded, or it will either miss a real leak or trigger a phantom one. The platform tracks each sensor's health, flags stale or missing samples, and quantifies how complete the coverage actually was over any period, so an operator can state with confidence that the site was genuinely watched and not silently blind. Alert thresholds then have to be set and tuned per site and per condition, because a level that is normal beside a loading rack would be alarming at a wellhead.
Reconciliation is where continuous monitoring becomes defensible rather than merely instrumented. When an alert fires, the record has to tie the anomaly to what the rest of the site was doing - a scheduled blowdown, a tank filling, a compressor trip - so permitted releases are distinguished from unintended leaks, and it has to capture the follow-up: who was dispatched, what they found, and when it was repaired. A cloud monitoring platform such as Merobix that already holds the process context alongside the emissions data can correlate the two automatically, close the loop from detection to repair, and produce the continuous compliance reporting that regulators expect from an approved alternative method. The sensors detect gas; the monitoring platform is what makes the program credible.
It can where a regulator approves it as an alternative test method or alternative work practice, which typically requires the operator to demonstrate that the continuous program detects and drives repair of leaks at least as effectively as the survey it replaces. Approval hinges on the whole system - sensor coverage, detection thresholds, investigation procedures, and record-keeping - not just the hardware. Many operators run continuous monitoring alongside a reduced survey cadence rather than eliminating surveys entirely.
An optical gas imaging survey is a periodic, human-operated inspection that localizes a leak to a specific component but only sees the site during the visit. A fixed sensor network watches continuously and detects site-level emissions between what would have been survey dates, but usually indicates that a site is leaking rather than pinpointing the exact flange. The two are complementary: continuous monitoring flags the site, and a targeted survey then locates the component for repair.
False alarms come from wind shifting a permitted plume onto a sensor, tanks breathing with temperature, scheduled blowdowns, and drifting or faulty instruments. They are reduced by fusing gas readings with wind data so the system knows which sources a sensor could plausibly see, by requiring multiple sensors to agree before escalating, and by reconciling alerts against known site activity so permitted releases are filtered out. Validating sensor health to exclude bad readings is equally important.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.