FAQ — frequently asked questions¶
General¶
What do I need to get started?¶
At least an annual-maxima series for one duration (ideally 5 durations and ≥ 15 years of data). If you don't have data, load one of the Calabria samples from the Info modal → Resources.
Can I use Runoff Lab without internet?¶
For operational use yes: the app runs in the browser and saves to localStorage. Only AI Import requires connectivity (calls a backend model).
Does my data stay private?¶
Yes. The study is saved in your browser (localStorage) and the local
.hgstudy file. Nothing is sent to the server except:
- AI Import sends the file to the backend (not retained);
- PDF export if it contains server-side generated images.
Save and crash recovery¶
I closed the tab without saving. Can I recover?¶
Yes. Reopen the app: browser autosave restores the session automatically.
For safety, save a .hgstudy with Ctrl+Shift+S: it's a portable JSON.
The "unsaved changes" badge stays on¶
It means the browser autosave is up-to-date but you haven't downloaded the
final .hgstudy for your most recent change. Press Ctrl+S (autosave) or
Ctrl+Shift+S (file download) to archive it.
Series and durations¶
How many years do I need?¶
- Below 10 years: very uncertain estimates; use only for exercises.
- 10-20 years: usable with care, prefer Gumbel-moments or L-moment methods.
- >20 years: all probabilistic families are comparable.
- >50 years: even TCEV level 0 (pointwise) becomes reliable.
Can I use durations other than the standard ones?¶
Yes. You can add custom durations in the study (e.g. 30′, 45′, 90′). All analyses use them if there are ≥ 2 values.
The 60' duration has 5 years and 24h has 30. What happens?¶
Each duration is analysed independently. The 60' will come out with high uncertainty, the 24h will be robust. The rainfall curves will have a good fit at long durations and less so at short ones — watch the \(R^2\).
Analyses¶
Which family should I pick?¶
Start with Gumbel-moments: simple, accepted by almost all authorities. Add a second family for comparison (GEV-Lmoments or TCEV-1) and document both in the final report.
Can I compare multiple analyses?¶
Yes. Add as many as you want in the same study: the page shows a side-by-side table with each method's \(h(T)\) values.
TCEV doesn't appear in the menu¶
The TCEV section is visible only with the interface in Italian: the VA.PI. dataset is specific to Italian regions. Switch the language to IT from the language menu.
Rainfall curves¶
My curve's \(R^2\) is 0.85 — what do I do?¶
The law \(h = a \cdot t^n\) doesn't fit well. Possible causes:
- Sparse series at some durations → unstable analysis.
- Climatic discontinuities or rain-gauge relocation.
- Mountain basin where short and long durations follow different regimes.
Try: TCEV level 3 (more parameters); or build separate curves for groups of durations (short vs long).
Which return period for design?¶
Typical Italian standards:
- Urban sewerage: T = 5–10 years.
- Minor culverts and crossings: T = 25–50 years.
- Hydraulic mitigation works, detention basins: T = 100–200 years.
- Settlement defence (PAI): T = 200 years and sometimes T = 500.
SCS-CN¶
What CN for a mixed area?¶
Use the CN wizard in the hydrograph panel: split the basin into sub-areas, assign soil type (A/B/C/D) and cover to each; the area-weighted CN is computed automatically.
How do I estimate Tlag?¶
Tlag wizard: enter A (km²), L (km, main channel), average slope i (%). You get \(T_c\) with 4 classical formulas (Kirpich, Giandotti, Pasini, Pezzoli) and the suggested \(T_{\text{lag}}\) as \(0.6 \cdot T_c\).
My basin is 500 km², does SCS-CN still work?¶
Marginally. SCS-CN works best under 250 km² with concentration times < 24 h. For larger basins use a distributed model (not implemented in Runoff Lab — use dedicated software like HEC-HMS).
AI Import¶
The AI made a mistake on one value¶
Open the preview → edit the cell directly → confirm. Manual corrections survive.
The PDF is poorly scanned — does it still work?¶
Generally yes, but inspect the preview carefully. If OCR quality is low, open the PDF, copy the text manually from the table, and paste in the modal's "free text" box.
How many credits does an import cost?¶
10 credits (\(\text{rainfall.import.ai}\)). Visible on the wallet chip before confirmation; rejected if the balance isn't enough.
Export¶
Can I export only one section?¶
For now the report is single and contains the whole study. If you want only the curves, delete the other sections before export — or trim the PDF afterwards.
The formulas look broken in the PDF¶
All formulas are rendered via MathJax in HTML and converted to SVG in the
PDF. If you see raw code (\frac{...} literal), there's a bug:
report it.
Common errors¶
"No valid item for the computation"¶
Analyses require ≥ 2 values per duration. Check in the Station section that at least two non-empty rows exist for each duration you want to use.
"Curves not available"¶
To create a curve you need at least one completed analysis. First go to the Analyses panel and add at least one.
"Hydrograph not available"¶
The SCS-CN hydrograph requires a synthetic hyetograph already built (or a direct curve + duration). Build a hyetograph first in its dedicated panel.