Humboldt County Sheriff | License plate readers

What the county bought, and with whose money

Seven cameras in the 2022-23 financial year, two more in 2023-24. Nine in total, and both staff reports name the funding source in the same words.

The first seven came from a Byrne JAG Local grant, which is a US Department of Justice programme, at a cost of $21,150 with one year of maintenance included. The next two came from a Homeland Security grant, which is federal money as well, at $14,000 with two years included.

The signed order form matches the staff report to the dollar. Seven “Falcon” cameras at $2,500.00 each a year, an annual recurring total of $17,500.00, two implementation charges, and a Year 1 subtotal of $21,150.00. The term was 24 months and the total contract amount was $38,650.00. An amendment signed in July 2025 renewed it through 30 June 2027.

There is one more line in the money that is worth reading twice. The 2024 report told the board that maintenance after the first term would be paid from “NorCal Coalition donation funding”. The 2026 report says what happened to that plan: the arrangement “did not occur due to budget cuts”. Upkeep now costs $21,000 a year and comes out of the General Fund, Prop 172, the COPS grant and a Homeland Security grant.

The policy is stronger than most, in two specific places

Two provisions in Policy 437 are better than the equivalent clauses in every other California plate reader policy this site has read.

The first is retention. Most of these policies set a minimum of one year. This one sets a maximum of 30 days:

All data is stored by Flock Safety in their secure cloud until 30 days, at which time it is purged. None of this data is stored by the HCSO, unless it has been deemed evidence.

The second is what it takes to run a search. The standard clause in this family of manual says that “Reasonable suspicion or probable cause is not required before using” the system, and this policy says that too. Then it adds two requirements the others do not have.

The first is that the search needs a reason, recorded with it:

However, there must be a sufficient investigative reason to conduct searches using the Flock Safety ALPR database

The second says what will do as a reason:

A case number or call for service (CFS) number is required to justify any search, and shall be entered into the database.

A case number is a real constraint. It ties every query to an open matter, and it leaves a trail that an audit can follow. The policy requires that audit monthly, which is more often than any other county documented here.

There is a third provision that goes further than anything else we have read. Every California plate reader policy in this family says data may not go to out-of-state or federal agencies. This one says where that rule has to reach:

Accordingly, SB 34 does not permit California LEAs to share ALPR information with private entities or out-of-state or federal agencies. This prohibition applies to ALPR database(s) that LEAs access through private or public vendors who maintain ALPR information collected from multiple databases and/or public agencies.

That last sentence is the one that matters. Sharing data is not usually how plate data leaves a county. Access is. A vendor holds the database, other agencies hold accounts, and the question is what the settings say. Humboldt’s policy is the only one on this site that names that mechanism.

Where it came from. The Department of Justice sent that paragraph to every law enforcement agency in California in October 2023, in Information Bulletin 2023-DLE-06. The policy reproduces it nearly word for word, citation included. That does not make it less notable, it makes it a different kind of notable: this office took the state’s written guidance and put it in the document that binds its own deputies. Most agencies did not.

And then there is section 437.3

The policy has a section titled FLOCK SAFETY TECHNOLOGIES. Here it is in full:

The HCSO exclusively contracts with Flock Safety for ALPR equipment, processing, maintenance, storage, and purging. Flock Safety is a company that provides security and surveillance technology, primarily focused on license plate recognition (LPR) cameras, live video cameras, drones, gunshot detection and data analytics. Their systems are used by neighborhoods, businesses, and law enforcement agencies to help deter crime and aid in the investigation process. Flock Safety technology reduces crime, recovers stolen vehicles, and improves community safety while maintaining a core focus on responsible data usage. Flock Safety technology is used to capture objective evidence without compromising on individual privacies.

The first three sentences describe a contract and a product line. The last two are claims about what the product does, in the present tense, with nothing cited.

Whether plate readers reduce crime is a live research question. The county’s own 2026 staff report says the evidence is “inconclusive” and describes “modest improvement in clearance rates for some property crimes near cameras, but no consistent evidence of broader crime reduction”. That sentence is in the same county’s records as the sentence above it.

A policy manual is not a brochure. It is the document that tells a deputy what is permitted, tells a supervisor what to audit, and tells a resident what the rules are. What it says about the vendor’s performance is not a rule, and it is the part of the document a reader has no way to check.

We checked one thing about it. The section is identical in the standalone policy PDF and in the full 955-page manual the office publishes, so it is not an artefact of one file. Section 437.4 carries the same pattern in a different form: in a list of the ALPR Administrator’s own duties, it states that “The Flock Safety ALPR system is crafted and maintained to be within compliance with all applicable laws.”

What the board did about it

On 2 June 2026 a supervisor put the programme back on the agenda. The item drew enough written comment to fill three separate attachment bundles in the packet.

Five motions were made. The record of each is precise, and it is worth setting out in full, because this is what a board actually deciding something looks like.

Carried, 4-1. Direct Public Works to return with a policy for Flock cameras on any county property, rights-of-way and easements for third-party applicants.

Carried, 4-1. Form an Ad Hoc of two supervisors, Sheriff’s Department representatives and a County Administrative Office representative to develop recommendations on surveillance use, and have Information Technology draft a countywide policy or ordinance from them.

Failed for lack of a second. Draft an ordinance prohibiting plate readers on county property, rights-of-way or facilities.

Failed, 2-3. Deallocate the $21,000 from the Sheriff’s budget by reducing the General Fund contribution to the Flock camera line.

Failed for lack of a second. Deny that allocation.

The board built oversight and declined to stop the programme or take its money. Both halves of that are the decision.

What the record does not say

The audits exist by policy, monthly, since at least May 2026. Not one result appears in anything we read.

The Sheriff’s Office publishes a link to a Flock Safety transparency portal on its Policies and Procedures page, alongside the policy manual and the immigration policy. That portal is the county’s own answer to a resident asking how to check. We could not read it. The vendor’s site returns a Cloudflare challenge to our fetcher, and a bot challenge is a gate rather than a request, so we stop there. Someone with a browser can open it, and that is the first item on this county’s retrieval list.

We also could not establish when the cameras were installed. The purchase years are in the county’s own financial disclosures. The statute that governs this requires the opportunity for public comment before an agency implements a programme, and the board took that item on 9 April 2024. The Sheriff’s report addressed the sequence directly, and its explanation belongs here:

The Sheriff’s Office has previously used an ALPR system within several contract cities and in the unincorporated areas of Humboldt County, prior to the law change. The ALPR system being implemented now consists of cameras mounted to street poles and/or other fixed points along the roadway

Purchase is not installation, and installation is not implementation. We do not know the dates that would connect them, so we are not drawing a conclusion from the ones we have. The question is in the records request on this site.