Note: Updated on 3/19/26 after receiving additional information from SJRA. It’s unlikely either of these projects will ever be built.

3/18/26 – The San Jacinto River Authority (SJRA) has released a 661-page “conceptual engineering feasibility study” on two flood-control dams in the Spring Creek watershed. One is on Birch Creek and the other on Walnut Creek. Both are dry detention dams that would capture floodwater temporarily and release it slowly after the peak of a storm passes.

The Walnut Creek reservoir could hold approximately 13,000 acre feet of stormwater and Birch could hold 9,000 acre feet. The 22,000 acre feet combined represent enough to reshape flood peaks significantly in the immediate area. Downstream areas would also benefit, but to a smaller degree.

The Benefit-Cost Ratio (BCR) on the Walnut Creek Dam barely meets Federal funding thresholds. However the Birch creek BCR is substantial. As a result, the study recommends pursuing both projects, but Birch Creek first.

The study also recommends:

• Another study to design dams on both creeks
• Refining property acquisition prices
• Pursuing the two projects independently rather than jointly.

That latter recommendation is because the combined BCR for the two dams falls far below 1.0 due because the areas benefitted include some duplication. So, they won’t get funded together. Costs far exceed benefits.

Finally, the feasibility study also recommends finding an entity willing to take ownership of the project and a funding mechanism to maintain the projects in Waller County in perpetuity! However, that section of the study does not list Waller County as a potential partner.

Net: While the tone of this study feels optimistic, many obstacles stand in the way that aren’t directly addressed on page 1. This feasibility study never does render an opinion on feasibility. It simply calls for another study to determine whether land can be acquired and what the real costs are.

Flood-Reduction Benefits of Dams

The Birch and Walnut Creek dams would have probable maximum inundation areas of 920 acres and 1,370 acres respectively; 640 acres and 940 in a hundred-year storm. If ChatGPT’s calculations are correct, they could shave peak flows near the creeks by a quarter to a third.

However, downstream at the confluence of Spring Creek and the West Fork, their impact would be much smaller – only 3.1% of the total flow. See below.

100-year Peak-Flow Reduction as a Percentage

Location	Existing 100-year discharge (cfs)	Walnut only	Birch only	Combined
On Walnut Creek	18,334	34.8%	24.0%	58.0%
Walnut Creek Confluence	48,330	2.5%	1.8%	4.0%
SH 249	46,808	7.3%	5.0%	11.6%
Kuykendahl	58,220	4.7%	3.3%	7.9%
Gosling	56,087	6.1%	3.5%	9.2%
I-45	60,814	5.5%	4.2%	7.7%
West Fork Confluence	69,337	1.9%	1.2%	3.1%

Upstream near the projects, peak reduction percentages are large because of the smaller drainage areas. But as you move downstream, inflows from other tributaries, such as Cypress Creek, dramatically reduce the percentages. You’re seeing the impact of much larger areas being drained.

Benefit/Cost Ratios

To meet Federal funding requirements, the benefits of a project must exceed its costs. And those benefits are typically calculated by the number of homes taken out of the 100-year floodplain – in this case 335, most of which are residential.

Estimating the value of those 335 structures then comparing them to the cost of the dams, shows that each project has a favorable Benefit/Cost Ratio. But Walnut Creek’s BCR of 1.05 just barely exceeds the Federal funding threshold of 1.0. Said another way, benefits barely exceed costs.

Worse, the reported BCRs for both projects include “social benefits.” The federal government no longer allows those as of 2025. But the study authors elected to keep them in the BCRs they reported. That’s because the projects can’t meet the 1.0 requirement without them. The study authors say on page 5 of the executive summary, “…these benefits are not being considered by FEMA at this time.” However, they add, FEMA may re-allow them in the future. (Bottom, Page 10 of PDF or 5 of Executive Summary.)

But that’s not the only B52-sized fly in the ointment. Because this study took so long, a giant solar farm grew up over and around the proposed Walnut Creek project area. Construction started in 2023 and completed in 2025. This study has been gestating since 2020. That drove up the projected purchase price of the Walnut Creek land and drove down the BCR.

Building the Walnut Creek project would require relocating approximately 880 acres of solar panels. That’s 1.375 square miles – 34% of the land in the total Walnut Creek Project.

The cost of relocating all those solar panels has driven up costs and driven down the BCR to the point where the benefits barely outweigh projected costs. The estimated ratio of benefits to costs is 1.05 – marginal.

Moreover, because the study took so long, the Birch Creek project is also endangered. According to the SJRA’s Matt Barrett, “That project would be more difficult to work around.”

But unlike Walnut Creek, the Birch Creek numbers apparently do not incorporate a workaround. So, it’s not totally clear how real the numbers below are.

Cost Per Structure Removed from Floodplain

Both dams together have an estimated total cost of $298 million.

That puts the estimated average cost per structure removed from the floodplain at $890,000.

And 42% of the housing in the project areas qualifies as low-to-moderate income (LMI). However, the entire 661-page report does not use the word “elevate” once. Nor does it use the word “buyout” once. Evidently, the study authors did not consider those alternative mitigation options. Both are classic FEMA strategies to reduce mitigation costs. And one Federal official I talked to said Federal dollars are available for both.

According to Barrett, “Elevation could potentially be a viable strategy in at least some locations/scenarios, but this study was focused on the feasibility of reservoirs in the Spring Creek watershed.

What Study Does/Does Not Show

The feasibility study covers topics such as probable costs, potential sources of funding, potential sponsors, land acquisition hurdles, environmental issues, permitting steps, probable designs, alternative dam locations, soils issues, cost-benefit analyses, and more.

However, even though the report is billed as a feasibility study, the conclusion does not state whether the proposed dams are feasible. It leaves that determination up to those who will debate the disparate findings.

Neither did I find discussions about:

• Cheaper mitigation options
• BCR calculations without social benefits included
• The likelihood of social benefits being re-included in the official formula
• What the cost of the proposed next study would be.

In fairness, “social costs” may sound fuzzy. But they include major real-world impacts of flooding, such as displacement, temporary housing, health impacts, economic disruption, school closures, tax losses, etc. So, real-world benefits likely exceed what the official formula allows.

The Costs of Not Taking Action

Also among topics I did not see in the report were the costs of not taking action. The projects are proposed for fast growing areas in far northeast Waller County.

Areas most positively impacted by these projects include Klein, Spring and the Woodlands. However, Humble, Kingwood and the Lake Houston area would benefit to a lesser degree. At the US59 bridge, the dams would reduce the height of a 100-year flood by an estimated 4 inches. That might not sound like much until the water starts creeping up your slab.

But peak reduction is only part of the story. The dams would also help keep peaks from other Lake Houston tributaries from stacking on top of each other and creating backwater effects.

Policy Implications

Upstream development makes the case for these dams more urgent, while also making delay more expensive. Why?

• Land gets more expensive
• More structures enter harm’s way
• More roads/utilities complicate acquisition and permitting
• Basin footprints become politically harder to preserve.

So, there is a race between:

• Locking in regional storage now, or
• Letting development consume the very geography needed for floodwater storage.

Delay makes the projects:

• More necessary
• More expensive
• Less effective

If Waller County urbanizes hard over the next decade, then the region may face a worse choice later:

• Buy much more expensive land for detention
• Widen channels downstream
• Dredge more often
• Rely more heavily on reservoir operations
• Or accept higher recurring damages.

These projects are not just flood-control projects; they are also land-preservation decisions. Ironically, Houston considered buying land in these same areas in 1985 when the price was a fraction of what it is today. But the rural land didn’t justify the BCR at the time.

Conclusions of Report

You can find the conclusions of the report on Page 47 of the study.

“One of the important next steps includes identifying a project sponsor within the region that will continue to move the projects forward,” says one of the conclusions.

It seems to me, that needs to happen before proceeding with design of the dams. Without someone willing to push the project forward, what’s the point of a final design that sits on the shelf for decades until it’s no longer doable?

Someone also needs to find whether the land can even still be purchased. If it’s already locked up, another study is a non-starter.

For More Information

The size of the entire study on the SJRA site is more than 330 megabytes. Several people have reported trouble downloading. So I have broken the study up into smaller chunks. See below.

SJRA Spring Creek Dams Feasibility Study – Evaluation of retention sites on Birch and Walnut Creeks. Entire file was 331 megs. Even when “reduced,” the 661 page report weighed in at 63 megs. So I broke it up into several sections to make it easier to download. I have also copied this information to the SJRA tab on my Reports page.

• Executive Summary – Overview, Table of Contents and Conclusions – 2.4 megs
• Exhibits 1 and 2 – 4.5 megs
• Appendix A1 – 5.1 megs
• Appendix A2 – 2.3 megs
• Appendix B1 – 3.2 megs
• Appendix B2 – 5.3 megs
• Appendix B3 – 3.8 megs
• Appendix C – 7.1 megs
• Appendix D – 1.7 megs
• Appendix D Maps – 28.7 megs
• Appendix E – 327 KB
• Appendix F – 380 KB

Posted by Bob Rehak on 3/18/2026

3123 Days since Harvey