Abstract:<p>Rainfall in mountainous watershed presents high spatial variability due to elevation effects, and this introduces uncertainty in forecasting hydrological hazards such as water floods and debris flows. This study investigated the spatial variation of rainfall in a small watershed with a network of 10 rain gauges. A rainfall-elevation relationship was established based on data from 52 rainstorm events, which provides a method for rainfall estimation within the watershed. Result indicated that lower errors of interpolation occur when the rainfall amount is high, and that it is more difficult to estimate rainfall in high-elevation regions. Rain gauges become less representative when the distance between gauges is >3.0 km. The spatial variation of rainfall suggests that the gauge at the lowest elevation, or a single gauge within the source region, shows non-negligible errors with regard to calculating water flood discharge and identifying rainfall thresholds for debris flows. This study contributes to the understanding of event rainfall distribution and its impact on hydrological hazard forecasting in a small mountainous watershed.</p>
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the impact of the spatial variability of rainfall on the prediction of hydrological disasters (such as floods and debris flows) in small semi - arid mountainous watersheds. Specifically, due to topographic effects, especially the influence of altitude, rainfall in mountainous watersheds exhibits a high degree of spatial variability, which brings uncertainty to the prediction of hydrological disasters. By analyzing the spatial variation of rainfall in a small watershed with a network of 10 rain gauges, this paper establishes a rainfall estimation method based on the rainfall - altitude relationship and evaluates the interpolation errors at different altitudes, rainfall amounts and rain gauge densities. In addition, the uncertainty when using different rainfall input data for flood and debris flow forecasts is also studied to determine the influence of the representativeness of rain gauges.
### Main research contents:
1. **Relationship between rainfall and altitude**:
- The study found that the rainfall amount increases with the increase of altitude. The relationship between the average event rainfall amount and the normalized altitude is established:
\[
R_m = 8.9e^{0.44h}, \quad R^2 = 0.7902
\]
where \( R_m \) is the average event rainfall amount (mm) and \( h \) is the normalized altitude.
- Further analysis shows that 75% of the events (39 out of 52) show an exponential relationship, which can be approximately described by the following formula:
\[
R = ae^{bh}
\]
where \( R \) is the rainfall amount (mm), \( h \) is the normalized altitude, the coefficient \( a \) ranges from 4.50 to 35.83, and the index \( b \) ranges from 0.255 to 0.857.
2. **Differences between rain gauges at similar altitudes**:
- Using the data of two pairs of rain gauges at similar altitudes (R4 and R5, R9 and R10), the correlation and differences of rainfall measurement values are analyzed. The results show that the rainfall data of these two pairs of rain gauges have a high correlation (Pearson correlation coefficients are 0.857 and 0.789 respectively), but there are certain deviations.
3. **Rainfall estimation based on altitude**:
- A rainfall estimation method based on altitude is proposed, which combines the relationship between rainfall amount and altitude:
\[
R = (0.69R_1 - 1.04)R^{-0.32h} e^{1.53h}
\]
where \( R \) is the rainfall amount (mm) at a certain point in the watershed, \( R_1 \) is the rainfall amount (mm) recorded by the lowest - altitude rain gauge R1, and \( h \) is the altitude relative to R1.
- The error of this method is evaluated, and it is found that when the rainfall amount of R1 is high, the error is low. If a 40% error is acceptable, then the rainfall amount of R1 should be greater than 15.0 mm and 17.5 mm respectively.
4. **Influence of rain gauge density and location on interpolation results**:
- By numerically generating rain gauge networks with different densities and locations, the influence of rain gauge density and location on interpolation results is studied. The results show that as the rain gauge density increases, the interpolation error gradually decreases.
- The interpolation error is small in low - altitude areas and large in high - altitude areas, indicating that it is more difficult to estimate rainfall in high - altitude areas.
- When the distance between rain gauges exceeds 3.0 km, the representativeness of rain gauges decreases significantly.
### Conclusion:
This study provides important insights for understanding the spatial distribution of rainfall in small mountainous watersheds and its impact on the prediction of hydrological disasters. By establishing an estimation method based on the rainfall - altitude relationship, the uncertainty of rainfall data in mountainous areas can be better dealt with, thereby improving the accuracy of hydrological disaster prediction.