Reynolds Number Calculator: Regime Check for Flow
Calculate Reynolds number (Re = ρvL/μ or Re = vL/ν) and classify flow as laminar, transitional, or turbulent. Supports internal pipe flow, external flow over flat plates or cylinders, and generic configurations. Compare up to 5 scenarios.
Reynolds number is the ratio of inertial forces to viscous forces in a flowing fluid. Re = ρvL/μ = vL/ν, where L is a characteristic length (pipe diameter for internal flow, position along a plate for boundary-layer flow, body diameter for flow around a cylinder). What it measures: does fluid momentum transport overpower fluid friction? Low Re means viscosity dominates and the flow stays orderly (laminar). High Re means inertia dominates and small disturbances amplify (turbulent). The transition isn't a single number, it's geometry-dependent. For internal pipe flow, Re < 2,300 is reliably laminar, Re > 4,000 reliably turbulent, with the band between as a transitional mess where pressure drop and heat transfer turn unpredictable.
Reference: Flow Regime Thresholds by Geometry
| Geometry | Laminar | Transition | Turbulent | Notes |
|---|---|---|---|---|
| Internal pipe flow | Re < 2,300 | 2,300 to 4,000 | Re > 4,000 | Smooth pipes; roughness lowers thresholds |
| Flat plate (external) | Re_x < 5×10⁵ | 5×10⁵ to 3×10⁶ | Re_x > 3×10⁶ | x = distance from leading edge |
| Cylinder (crossflow) | Re_D < 2×10⁵ | 2×10⁵ to 5×10⁵ | Re_D > 5×10⁵ | D = cylinder diameter; wake dominates drag |
| Sphere | Re_D < 2×10⁵ | 2×10⁵ to 4×10⁵ | Re_D > 4×10⁵ | Drag crisis at transition |
| Non-circular duct | Re_Dh < 2,300 | 2,300 to 4,000 | Re_Dh > 4,000 | Use hydraulic diameter D_h = 4A/P |
Identifying Flow Regime First (Laminar, Turbulent, Transitional)
The first job of the Reynolds number is to settle which regime you're in. Re = ρVL/μ compares inertial forces (ρV²) to viscous forces (μV/L). When viscous forces dominate (low Re), fluid layers slide past each other without mixing, that's laminar flow. When inertial forces dominate (high Re), eddies form and the flow becomes chaotic, turbulent. The transition isn't instantaneous; there's an unstable zone where small disturbances can trigger turbulence.
Why thresholds matter for design: laminar flow gives predictable pressure drop (ΔP ∝ V) but poor heat and mass transfer. Turbulent flow gives better mixing and higher heat transfer coefficients, but ΔP scales with V to the power 1.75 to 2. Pumping power scales with ΔP × Q, so turbulent flow costs more energy. If you design a heat exchanger assuming turbulent correlations and operate in laminar, the Nusselt number drops by 50 to 80 percent, and your exchanger undersizes drastically.
Threshold variability: published values (Re_crit ≈ 2,300 for pipes) assume smooth walls, fully developed flow, and no external disturbances. Surface roughness, entrance effects, vibration, and pulsating flow all lower the critical Re. In some labs, pipe flow has stayed laminar to Re = 40,000 by eliminating disturbances. In industrial practice, assume transition starts at Re ≈ 2,000 for safety.
For fully developed flow in circular pipes, the classic thresholds are: laminar for Re < 2,300, transitional for 2,300 ≤ Re ≤ 4,000, and turbulent for Re > 4,000. These values trace back to Osborne Reynolds' 1883 dye-injection experiments and remain valid for smooth pipes with undisturbed inlets. Friction factor selection follows: in laminar flow, Darcy friction factor f = 64/Re exactly. In turbulent flow, use Colebrook-White or Swamee-Jain equations that depend on both Re and relative roughness ε/D. The transition region is unpredictable, so most engineers either avoid it or interpolate cautiously.
Watch out: high-viscosity fluids (motor oils, polymer melts, slurries) often give Re < 100 even at high velocities. Don't assume turbulence exists, check Re explicitly. Conversely, water in typical industrial pipes almost always runs turbulent (Re > 10,000) unless velocities are very low.
Choosing Geometry: Pipe vs. Plate vs. Cylinder vs. Open Channel
The characteristic length L in Re changes by geometry, and picking the wrong one is the single biggest source of error in flow regime classification. For internal pipe flow, L is the bore diameter D. For non-circular ducts, L is the hydraulic diameter D_h = 4A/P (with A as cross-sectional area and P as wetted perimeter). For flat-plate boundary layers, L is the distance x from the leading edge, so the local Re_x grows with position. For external flow over a cylinder or a sphere, L is the body diameter. For open-channel flow, L is also the hydraulic diameter, defined the same way as for pipes but with the free surface excluded from the wetted perimeter.
For internal flow, fully developed pipe flow has a sharply defined transition near Re = 2,300 to 4,000. Boundary-layer flow on a flat plate transitions much later, near Re_x ≈ 5×10⁵. Crossflow around a cylinder transitions in the boundary layer near Re_D ≈ 2×10⁵, and the wake transitions earlier (Re ≈ 40 for vortex shedding to begin). These thresholds aren't arbitrary, they reflect the different stability properties of the underlying flow.
Hydraulic diameter for non-circular ducts:
- Circular pipe: D_h = D (wetted perimeter πD, area πD²/4)
- Rectangular duct (a × b): D_h = 2ab/(a+b). HVAC duct sizing leans on this.
- Annulus (D_o, D_i): D_h = D_o − D_i. Concentric tube heat exchangers use this.
- Square duct (side a): D_h = a
Hydraulic diameter works well when the duct shape is roughly circular (aspect ratio under 4:1). For very elongated shapes (slot flows, thin gaps in plate-and-frame heat exchangers), secondary flows and corner effects make D_h correlations less accurate. Specialized correlations exist for high-aspect-ratio ducts, and Munson's Fundamentals of Fluid Mechanics tabulates them.
For external flow over a flat plate, the local Reynolds number Re_x = Vx/ν increases with distance x. Near the leading edge, Re_x is small and the boundary layer is laminar. As x increases, Re_x eventually exceeds about 5×10⁵, and the boundary layer transitions to turbulent. The exact transition location depends on freestream turbulence, surface roughness, and pressure gradient. Aircraft and automotive designers shape wings and bodies to delay transition because laminar flow reduces drag. Natural laminar flow airfoils maintain favorable pressure gradients as long as possible. Surface imperfections (rivets, bug strikes, paint chips) trigger early transition and negate the drag benefit.
Cross flow around a cylinder produces dramatically different wake patterns depending on Re_D. At Re < 5, flow is symmetric and attached. Between Re = 5 and Re = 40, steady separation with twin vortices forms behind the cylinder. From Re = 40 to Re ≈ 2×10⁵, vortices shed alternately (Kármán vortex street) at a frequency given by Strouhal number St ≈ 0.2. Above Re ≈ 3×10⁵, the boundary layer transitions to turbulent before separation, delaying separation and dropping drag sharply (the drag crisis).
Boundary Conditions That Change the Answer
Re tells you what the bulk flow wants to do. The boundary conditions tell you what actually happens at any specific point. Flow entering a pipe from a tank or a fitting needs distance to develop into a fully-developed velocity profile. In the entrance region, velocity near the wall accelerates while the core decelerates, creating higher wall shear and pressure drop than fully-developed flow. The hydrodynamic entrance length L_e is the distance where the boundary layer fills the pipe.
Entrance length formulas:
- Laminar: L_e/D ≈ 0.05 Re. At Re = 2,000, L_e ≈ 100 D.
- Turbulent: L_e/D ≈ 4.4 Re^(1/6). At Re = 10,000, L_e ≈ 20 D.
In laminar flow at Re = 2,000, entrance length is 100 diameters. For a 1 inch PVC schedule 40 line, that's over 8 feet of straight pipe before you can use fully-developed correlations. Most laminar flow in short industrial piping is actually developing flow with higher pressure drop than the textbook formula predicts. If you're sizing a pump for a viscous polymer line of moderate length, multiply the predicted ΔP by 1.5 to 2 as a sanity factor.
Surface roughness changes the answer too. Drawn copper at ε ≈ 0.0015 mm and clean PVC at ε ≈ 0.0015 mm both behave hydraulically smooth at typical service Re. Commercial steel at ε ≈ 0.046 mm starts showing roughness effects around Re > 10⁵. Cast iron with ε ≈ 0.26 mm and old corroded carbon steel with ε up to 1 mm push the Moody chart firmly into the fully rough regime where the friction factor depends only on ε/D and not on Re at all.
Inlet conditions on aircraft wings, sails, and turbine blades also matter. A wing in clean atmospheric air sees lower freestream turbulence than the same wing in wake of another aircraft, and that changes the effective transition Re. Wind-tunnel testing has to match the freestream turbulence intensity of the operating condition or the boundary-layer behavior won't scale.
Practical tip: instrument taps, flowmeters, and control valves should sit downstream of entrance effects, typically 10D minimum for turbulent flow, 50D or more for laminar. ASME MFC-3M gives specific upstream straight-run requirements for orifice and Venturi installations.
Gauge vs. Absolute Pressure: Why It Matters in Engineering
Reynolds number itself doesn't care about gauge versus absolute pressure, because Re is built from velocity, length, density, and viscosity, not pressure. But the moment you compute Re for a real pipe and want to translate the result into pressure drop or pump sizing, the gauge versus absolute distinction matters.
Pressure drop calculations using Darcy-Weisbach (turbulent) or Hagen-Poiseuille (laminar) give a ΔP that is the same number whether you express it in gauge or absolute. So those calculations are unit-neutral. The trap shows up at the boundaries. NPSH (net positive suction head) calculations on a pump need absolute pressure at the inlet, because the cavitation threshold is the absolute vapor pressure of the fluid. A 0.0946 m³/s line of crude oil at 0.5 m/s through a 4 inch schedule 40 pipe has Re around 5,000 to 10,000 depending on temperature, and that determines f. But the pump margin against cavitation depends on absolute pressure at the pump suction.
Gas density used in Re also depends on absolute pressure. For natural gas in a pipeline, ρ scales linearly with absolute pressure at constant temperature. So a Re calculation at 7 MPa absolute uses ρ ≈ 50 to 60 kg/m³ for a typical gas. The same line at 100 kPa absolute (atmospheric) would have ρ closer to 0.7 kg/m³, a factor of 70 difference. That changes Re by the same factor and absolutely changes the regime classification. If you accidentally use gauge pressure (so 6.9 MPa instead of 7 MPa absolute) the answer is barely off. If you use gauge instead of absolute on a low-pressure system, the answer is wildly wrong.
The practical rule: use whatever pressure units your calculation expects, but write the unit explicitly on the page. Most fluid property tables (NIST WebBook, REFPROP) list properties at absolute conditions, so when you look up viscosity at 20°C and 101 kPa, that 101 kPa is absolute. Mix this up with a gauge reading from a pressure transmitter and you can shift Re by orders of magnitude on a gas line.
Conservation Laws Applied to the Specific Flow
Reynolds number falls out of the Navier-Stokes equations when you non-dimensionalise. Inertial terms scale with ρV²/L, viscous terms scale with μV/L². The ratio ρVL/μ tells you which set of terms governs. That's the conservation-of-momentum statement at the heart of Re. Mass conservation (continuity) and energy conservation enter when you actually solve the flow.
Common fluid viscosities at 25°C, from NIST WebBook:
- Water: μ = 0.89 mPa·s, ρ = 997 kg/m³
- Air (1 atm): μ = 0.0185 mPa·s, ρ = 1.18 kg/m³
- Ethanol: μ = 1.07 mPa·s, ρ = 785 kg/m³
- Motor oil (SAE 30): μ ≈ 200 mPa·s, ρ ≈ 880 kg/m³
- Glycerin: μ ≈ 950 mPa·s, ρ = 1,261 kg/m³
Liquid viscosity drops sharply with temperature. Water runs from 1.79 mPa·s at 0°C down to 0.28 mPa·s at 100°C, a factor of 6 change. Gas viscosity instead rises with temperature, opposite to liquids. Using room-temperature viscosity for a hot process or a cold start can shift Re by a factor of 2 to 10 and flip the regime entirely. Frank White's Fluid Mechanics has accurate Sutherland correlations for gas viscosity-temperature, and Munson's text covers liquid temperature dependence.
Vortex-induced vibration (VIV) is a momentum-conservation phenomenon driven by Re. Above Re_D ≈ 40 in cross-flow over a cylinder, vortices shed alternately at Strouhal frequency f = St·V/D with St ≈ 0.2. The cylinder feels an oscillating side force at that frequency. When the shedding frequency matches a structural natural frequency, resonance can drive catastrophic oscillation. Tacoma Narrows in 1940 is the famous lesson, and modern offshore drilling risers use strakes or fairings to break up coherent vortex shedding.
Energy conservation appears in the heat transfer correlations. Laminar flow has fixed Nu (Nu = 4.36 for constant heat flux in fully developed pipe flow, 3.66 for constant wall temperature). Turbulent flow uses Dittus-Boelter Nu = 0.023 Re^0.8 Pr^n, where the Re^0.8 dependence shows just how strong turbulence is at moving heat. Going from Re = 1,000 (laminar) to Re = 10,000 (turbulent) increases Nu by roughly a factor of 5 to 10 in pipe flow, which is why heat exchanger designers push hard for turbulent flow in shell-and-tube designs.
For tube banks in crossflow heat exchangers (HVAC duct coils, oil coolers, condensers), the Re used in correlations is based on the maximum velocity through the bank, not the freestream. The drag crisis at Re > 2×10⁵ is rarely reached in process equipment, so most tube bank correlations assume subcritical flow with periodic vortex shedding.
Worked Example: Water at 20°C in a 50 mm Copper Pipe
Problem: water at 20°C flows through a 50 mm bore type L drawn copper pipe at average velocity 0.5 m/s. Properties at 20°C: ρ = 998 kg/m³, μ = 1.002 × 10⁻³ Pa·s. Determine the Reynolds number, classify the flow regime for the circular pipe, and compare with the transition Re for a flat-plate boundary layer.
Step 1: Plug into Re = ρvD/μ
Re = (998 × 0.5 × 0.050) / (1.002 × 10⁻³)
Re = 24.95 / 1.002 × 10⁻³
Re ≈ 24,950
Step 2: Apply circular-pipe regime thresholds
Re < 2,300 (laminar), 2,300 ≤ Re ≤ 4,000 (transitional), Re > 4,000 (turbulent). With Re ≈ 24,950, the flow is solidly turbulent, well past the transition band.
Step 3: Compare against flat-plate boundary layer
For external flow over a flat plate, transition kicks in around Re_x ≈ 5×10⁵, where Re_x = Vx/ν. To reach that on a flat plate at the same V = 0.5 m/s in 20°C water (ν = μ/ρ ≈ 1.004 × 10⁻⁶ m²/s):
x = Re_x × ν / V = 5×10⁵ × 1.004×10⁻⁶ / 0.5 ≈ 1.0 m from the leading edge.
So the same fluid at the same speed transitions to turbulent in a circular pipe almost immediately (Re=4,000 corresponds to D ≈ 8 mm), but takes about 1 m of plate length to transition externally. That's a factor of 200 difference in characteristic length, and it's why boundary-layer transition Re for plates lives at 10⁵ to 10⁶ while pipe flow transition lives at 10³.
Step 4: Pick a friction factor
Drawn copper has ε ≈ 0.0015 mm. ε/D = 0.0015/50 = 3×10⁻⁵, near hydraulically smooth. Swamee-Jain at Re ≈ 25,000 and ε/D = 3×10⁻⁵ gives f ≈ 0.0245.
Result:
Re ≈ 24,950 puts a 50 mm copper line at 0.5 m/s comfortably turbulent. Friction factor f ≈ 0.0245 from Swamee-Jain. Use the Darcy-Weisbach equation for pressure drop: ΔP/L = f × (ρV²/2)/D = 0.0245 × (998 × 0.25/2)/0.050 ≈ 61 Pa/m. Over a 50 m run, that's about 3 kPa total ΔP, easy duty for any small recirculation pump.
References
The Reynolds-number framework on this page assumes Newtonian fluids, steady incompressible flow, smooth walls, and single phase. Non-Newtonian fluids (polymers, slurries, blood) need a generalised Reynolds number or power-law form, because viscosity depends on shear rate. Pulsating flow, compressible gases at high Mach, and transient startup conditions can behave differently from steady-state correlations. Surface roughness lowers critical Re and changes turbulent friction factors, so always pull ε from the actual pipe specification (commercial steel ε = 0.046 mm, drawn copper ε = 0.0015 mm, PVC schedule 40 ε ≈ 0.0015 mm). Multiphase flows (gas-liquid, solid-liquid) have complex regime maps not captured by single-phase Re alone.
Limitations recap
- Newtonian assumption: viscosity independent of shear rate.
- Steady, incompressible flow: high-Mach gas and pulsating systems need separate treatment.
- Smooth surfaces: rough pipes shift transition and friction.
- Single-phase: gas-liquid mixtures need a multiphase regime map.
- White, F. M. (2016). Fluid Mechanics (8th ed.). McGraw-Hill. Standard reference for Reynolds number, transition thresholds, and flow correlations.
- Munson, Young, & Okiishi (2021). Fundamentals of Fluid Mechanics (8th ed.). Wiley. Coverage of pipe flow, boundary layers, and external flows.
- Schlichting & Gersten (2017). Boundary-Layer Theory (9th ed.). Springer. Advanced treatment of transition, stability, and turbulence onset.
- Reynolds, O. (1883). An experimental investigation of the circumstances which determine whether the motion of water shall be direct or sinuous. Philosophical Transactions of the Royal Society. The original dye-injection paper.
- Moody, L. F. (1944). Friction factors for pipe flow. Transactions of the ASME, 66(8), 671-684. The diagram-on-a-page that turned the Colebrook implicit equation into a chart engineers can read in 10 seconds. Still printed in the back of every fluid-mechanics text.
- Crane Co. (2018). Flow of Fluids Through Valves, Fittings, and Pipe (Technical Paper No. 410). The hydraulic-engineering reference for K-factors of valves and fittings, equivalent-length tabulations, and friction-factor lookups used in real piping-system pressure-drop calculations.
- NIST Chemistry WebBook: webbook.nist.gov. Authoritative source for fluid thermophysical properties.
- Engineering Toolbox: engineeringtoolbox.com. Quick reference for fluid properties and Reynolds number tables.
Troubleshooting Reynolds Number and Flow Regime Classification
Real questions from engineers and students stuck on transition zone predictions, hydraulic diameter calculations, temperature effects on viscosity, and why measured pressure drop doesn't match predicted values.
How do you tell if flow is laminar or turbulent?
Compute the Reynolds number Re = ρvD/μ, where ρ is fluid density, v is mean velocity, D is the characteristic length (pipe diameter for internal flow), and μ is dynamic viscosity. Below Re ≈ 2300 in a smooth round pipe, flow is laminar. Above Re ≈ 4000, it's fully turbulent. Between those, you're in the transition zone, where flow can flip-flop. The Reynolds number is dimensionless. It compares inertial forces (pushing fluid forward in straight lines) to viscous forces (smearing motion out). Small Re means viscosity wins, so fluid layers slide smoothly past each other. Large Re means inertia wins, and the flow breaks up into eddies and chaotic mixing. Water at 20°C (ρ = 998 kg/m³, μ = 1.0 × 10⁻³ Pa·s) flowing at 1 m/s through a 25 mm pipe gives Re = 998 · 1 · 0.025 / 0.001 ≈ 25,000, which is firmly turbulent. Drop the velocity to 0.05 m/s and Re falls to 1250, well into the laminar zone. Pipe surface roughness shifts the turbulent transition slightly but barely affects the laminar-turbulent threshold. The choice matters because pressure drop scales as v in laminar flow but roughly as v² in turbulent flow. A laminar correlation applied to turbulent flow underestimates pressure loss by 30-50%. For non-pipe geometries, the characteristic length changes. Use hydraulic diameter for non-circular ducts. For flow around immersed bodies, use the body's leading dimension. The Re framework still applies.
My heat exchanger design assumes turbulent flow for good heat transfer, but at cold startup the oil is so viscous that Re drops below 2,000. Should I redesign?
Not necessarily—but you need startup procedures. High-viscosity cold oil means laminar flow and heat transfer coefficients 50–80% lower than turbulent. Options: (1) preheat oil before startup, (2) oversize the exchanger for cold conditions, (3) accept slower warmup, or (4) recirculate through a heater before full production. Many refineries use startup heaters specifically for this. Calculate Re at both extremes (cold start, hot steady-state) and design for the limiting case.
I calculated Re = 3,500 for our pipe flow and used laminar correlations—now the measured pressure drop is 40% higher than predicted. What went wrong?
You're in the transition zone (2,300–4,000), which is notoriously unpredictable. At Re = 3,500, flow can be partially turbulent with intermittent bursts. Laminar correlations (f = 64/Re) underpredict friction; turbulent correlations (Colebrook-White) may overpredict. The safest approach: avoid operating in transition. If you must, use turbulent correlations with conservative margin—they'll overpredict slightly rather than dangerously underpredict.
I'm using water at Re = 50,000 in a pipe, but the textbook says critical Re is 2,300. Why isn't my flow transitioning to laminar?
You've got it backward. Re = 2,300 is the lower threshold—below it, flow is laminar. Above Re ≈ 4,000, flow is turbulent. At Re = 50,000, you're well into turbulent territory, and it would take extreme measures (eliminating all disturbances, vibration-free piping, smooth walls) to maintain laminar flow at that Re. In practice, once turbulent, flow stays turbulent unless velocity drops dramatically.
Our HVAC duct is rectangular (24" × 12"). I calculated Re using the long dimension but my coworker used hydraulic diameter. Who's right?
Your coworker is right. For non-circular ducts, always use hydraulic diameter: D_h = 4A/P = 4(24×12)/(2×24+2×12) = 1152/72 = 16 inches. Using the long dimension (24") overestimates Re by 50%, potentially leading you to expect more turbulence than exists. Hydraulic diameter collapses rectangular ducts to an equivalent circular pipe for friction and heat transfer correlations.
I measured flow rate with an orifice plate, then calculated Re to verify I'm turbulent—but my orifice correlation also needs Re. Is this circular?
Yes, but it converges quickly. Start by assuming turbulent flow (valid for most industrial orifices), use the turbulent discharge coefficient (C_d ≈ 0.61), calculate flow rate and velocity, then check Re. If Re > 10,000, your assumption was valid. If Re is lower, iterate with a corrected C_d from your orifice calibration curve. Most practical orifice meters run Re > 20,000, so one iteration is usually enough.
My professor marked me wrong for saying a golf ball dimple pattern trips turbulent flow at lower Re. Isn't that how dimples work?
You're partially right—dimples do trip turbulence, but that's not the main mechanism. Dimples energize the boundary layer, keeping it attached longer around the ball. This delays separation, shrinks the wake, and drastically reduces pressure drag. The boundary layer becomes turbulent earlier, yes, but turbulent boundary layers resist separation better than laminar ones. So dimples cause higher skin friction but much lower pressure drag—net drag reduction of 50%. Your professor may want the full picture.
We're scaling a ship model in a towing tank. The ship runs at Re = 10⁹ but our model can only reach Re = 10⁶. How do we compensate?
This is classic Reynolds scaling problem. At Re = 10⁶, your model's boundary layer is partially laminar while the full-scale ship is fully turbulent. Solutions: (1) trip turbulence artificially using sandpaper or trip wires near the bow, (2) use larger models to reach higher Re, (3) test at multiple Re and extrapolate, (4) accept that model tests underpredict friction and use correction factors (Froude scaling for waves, ITTC correlation line for friction). Naval architects have standard procedures for this—consult ITTC guidelines.
I looked up water viscosity at 20°C (1.0 mPa·s) but then used 60°C water in my process. My Re calculation was way off—how much does temperature matter?
Temperature matters enormously for liquids. Water viscosity drops from 1.0 mPa·s at 20°C to 0.47 mPa·s at 60°C—more than halving. This doubles your Re. If you calculated Re = 2,000 at 20°C, actual Re at 60°C is ~4,200—you jumped from transitional to turbulent. Always use viscosity at actual operating temperature. For oils, the effect is even more dramatic: SAE 30 oil can change viscosity by 10× over a 50°C range.
Our flowmeter manufacturer says to install it 10D downstream of elbows, but at our flow rate the entrance length is 50D. Which is right?
Both can be right for different flow regimes. The 10D rule applies to turbulent flow (Re > 10,000), where the entrance length L_e/D ≈ 4.4·Re^(1/6) ≈ 20D. In laminar flow, L_e/D ≈ 0.05·Re—at Re = 1,000, that's 50D. Your flowmeter spec assumes turbulent; if you're running laminar, you need more straight pipe. Also, flowmeter specs are for accuracy (eliminating swirl and velocity profile distortion), not just fully-developed flow. Check your actual Re before installation.
My Re calculation gives 2.5×10⁶ for air over a car at highway speed. The textbook says transition is at 5×10⁵—so the whole car is turbulent?
Not the whole car—the front is laminar. Re_x = Vx/ν increases with distance x from the nose. At highway speed (~30 m/s) in air (ν ≈ 1.5×10⁻⁵ m²/s), transition occurs at x ≈ 5×10⁵ × 1.5×10⁻⁵ / 30 ≈ 0.25 m—about 10 inches from the front. Beyond that, the boundary layer is turbulent. Your Re = 2.5×10⁶ probably uses total car length, which gives the Re at the rear. The front 10 inches sees laminar flow; the rest is turbulent.