--- name: landing-and-go-around description: Assess a Cirrus SF50 Vision Jet or SR22 landing and whether a go-around, balked landing or missed approach is possible — landing distance against LDA, runway contamination and RwyCC (SF50) or wet grass (SR22), Vref and the AFM go-around climb gradient (SF50) or the POH balked landing climb (SR22), and the missed-approach climb versus the required gradient, terrain and obstacles for each approach. Use for questions like "is a go-around possible from KASE runway 15 tomorrow at 2pm?", SF50 landing distance on a wet or snowy runway, "how far do I roll landing my SR22 on wet grass?", "can my SR22 make the missed-approach climb gradient?", or whether the SF50 can make a missed-approach climb gradient. Combines landing_performance, find_procedures, get_procedure, climb_profile, query_terrain and query_obstacles. --- # Landing and go-around ## Safety - Planning aid only. Figures from these tools are not approved by the FAA, the aircraft manufacturer or any supplement holder and are not for cockpit use; say so whenever you report them. - Never invent, estimate or interpolate a figure the tools did not return. If something is missing, name it and the input that would supply it. - Pass through every non-`value` status (`uncertain`, `offscale_high`, `offscale_low`, `not_available`, `not_authorized`, `invalid`), every warning and any `partial` completeness. Never replace them with a guess. - List every input with its source and time: `nasr`, `metar`, `taf`, `open_meteo`, `override`, `isa` or `assumed`, plus NOTAM numbers and data cycles. ## Inputs to collect - `aircraft`, an object chosen by `type` (ask if unknown; never assume): - SF50: `{"type": "sf50", "generation": "g2"}`, with `generation` one of `g1`, `g2`, `g2_updated_thrust`, `g2_plus` or `g3`. - SR22 (pre-G3 wing): `{"type": "sr22", "hartzell_propeller": false, "turbonormalized": false}`. Ask whether the aircraft has the Hartzell composite propeller supplement and the Tornado Alley turbonormalizing supplement; both are required. - The **landing** weight (`weight_lb`), not the takeoff weight. The tool refuses weights above the aircraft's maximum. - `airport`, `runway` and the arrival time, converted by performance-weather-for-time. - `flaps`: `100` by default. The SR22 has only `100`. The SF50 also takes `50` and `up`, and `50_ice` or `up_ice` when ice protection is on. Any other setting is refused with `invalid_argument`. - NOTAM inputs from notams-to-performance-inputs: `available_landing_distance_ft`, `contamination`, `contamination_depth_in`, `rwycc`. The SF50 takes every contamination type. The SR22 takes only wet grass: `contamination: "wet"` with `surface: "turf"`. - `vref_additive_kt`, only for an aircraft with a Vref (the SF50). It is refused without one. ## Steps 1. **Landing.** Call `landing_performance`: ```json {"aircraft": {"type": "sf50", "generation": "g2"}, "weight_lb": 5400, "airport": "KASE", "runway": "15", "time": "2026-11-05T21:00:00Z"} ``` Read `landing_run_ft`, `landing_distance_ft` (from 50 ft), the breakdowns and `runway_margins.landing_distance`, and the outputs the aircraft's `capabilities` give: - SF50 (`vref`, `go_around_climb_gradient`): `vref_kt` and `meets_go_around_climb_gradient`. That field is the AFM go-around gradient check at these conditions. If it is false or not a `value`, say a go-around is not established by the AFM data, and continue only to show the missed-approach picture. - SR22 (`balked_landing_climb`): `balked_landing_climb_gradient_ft_per_nm` and `balked_landing_climb_rate_ft_per_min`, the POH balked landing climb at these conditions. Use them in place of the go-around check. The SR22 POH has no Vref and no go-around gradient check: `vref_kt` and `meets_go_around_climb_gradient` are listed in `unavailable_outputs` and absent. Never supply a Vref. 2. **Approaches.** Call `find_procedures` with `airport` and `procedure_type: "approach"`. Call it without `runway` as well: circling approaches list no runway (`runways: []`, `landing: "circle_to_land"`), yet they may be how you reach runway 15. If the result warns of CIFP parse errors, say approaches may be missing. 3. **Missed approach.** For each approach, call `get_procedure` with `airport` and `identifier` (such as `RNV-F`). - The missed-approach point is the leg with `is_missed_approach_point: true`; note its `position` and `altitude_constraint`. - `missed_approach_legs` give the path: `CA` (course to an altitude), `DF` and `TF` (to a fix), `HM` (the hold). Courses are `reference: "magnetic"`. - `obstacle_clearance_assessed: false` means the server did not assess the procedure. 4. **Required gradient.** CIFP does not carry chart notes. The standard missed-approach climb is **200 ft/NM** unless the chart publishes a higher one. Ask the user for the chart's figure, or state that you used 200 ft/NM and that a published higher gradient replaces it. A NOTAM can also change it. 5. **Course in true.** Missed-approach course true = magnetic + east variation, taking the variation from a runway end in `get_airport` as `heading_true_deg − heading_magnetic_deg`. 6. **Climb.** Call `climb_profile` with the landing weight, `airport`, `runway`, `time`, `course_true_deg` (the initial missed-approach course in true) and `top_altitude_ft` (the highest missed-approach altitude constraint, usually the hold). - The tool has no go-around configuration. The `takeoff` schedule (SF50 91 KIAS) is the closest proxy for the initial climb; the SF50's `enroute_obstacle` (120 KIAS) applies later. The SR22 has `takeoff` and `enroute` schedules only. Say which you used. - For the SR22, compare the balked landing climb from step 1 with the required gradient as well. - Use `schedules.takeoff.minimum_ground_gradient_ft_per_nm`. A headwind on the missed-approach course raises the ground gradient; a tailwind lowers it. The `levels` array gives gradients per 1,000 ft if the requirement applies only up to an altitude. - `bands_without_gradient` and non-`value` statuses mean the manual's data does not cover part of the climb. The ice schedules are often offscale on a warm day; report that, but it does not affect the non-ice verdict. 7. **Terrain.** Call `query_terrain`: - `route`: the missed-approach point, then the fix positions along the missed approach. For a `CA` leg, project the point where the altitude is reached with `compute_geometry` (`operation: "destination"`, `course_true_deg`, `distance_nm` = altitude to gain ÷ gradient). At most 20 points. - `half_width_nm`: up to 2. The corridor is not a TERPS surface. - `starting_altitude_ft`: the altitude at the missed-approach point (the DA or MDA when the user gives it, otherwise the CIFP altitude there). - `gradient_ft_per_nm`: the climb profile's ground gradient, rounded down. - `clearance_ft`: state your choice; use 0 for a bare terrain comparison. - `vertical_datum`: `EGM2008`. Charted MSL altitudes are close to EGM2008 heights in the U.S. but not identical; treat margins of a few tens of feet as inconclusive. - `climb.status` is `conditional_clearance_met` or `not_established`. The latter is inconclusive, not proof of a conflict. Report `limiting.minimum_margin_ft`, its position, and any coverage gaps. 8. **Obstacles.** Call `query_obstacles` with `procedure_id` (the `id` from `find_procedures`) and a `radius_nm`, or with a `route` along the missed approach. Compare each obstacle's `elevation_msl_ft` with the altitude the climb gives at its distance (starting altitude + gradient × distance). ## Verdict per approach | Approach | Required (source) | Ground gradient (schedule) | Margin | Terrain | Obstacles | NOTAMs | Verdict | | --- | --- | --- | --- | --- | --- | --- | --- | | RNAV (GPS)-F | 200 ft/NM (standard; chart not checked) | `` ft/NM (takeoff) | `` ft/NM | `conditional_clearance_met` | `` | `` | Meets on these inputs | Use only "meets on these inputs", "does not meet" or "not established". Never call a go-around safe or certified. ## Answer format 1. Landing: landing distance against LDA with margin, contamination effect, and for the SF50 Vref and the AFM go-around check, or for the SR22 the balked landing climb gradient and rate. Note any `unavailable_outputs`. 2. The per-approach verdict table. 3. Inputs with origin and time; NOTAMs applied and uncertain. 4. The disclaimers from the results, including the manuals' configuration notes, and the aircraft's `documents`.