You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
## Albatross: A Climate Service for Seasonal Forecasting
20
+
**Albatross: A Climate Service for Seasonal Forecasting**
21
21
22
22
*Albatross is a bird designed for seasonal forecasting of hydroclimatic variables.*
23
23
24
24
Albatross is deployed as a **Web Processing Service (WPS)**, providing a robust, accessible interface for climate analysis. It produces deterministic forecasts of seasonal hydroclimatic variables (e.g., precipitation, temperature, streamflows,...) for any given location by leveraging global teleconnections.
25
25
26
26
It is built upon the Niño Index Phase Analysis (NIPA) framework by Zimmerman et al. (2016).
27
27
28
-
### ⚡ WPS Interface Overview
28
+
**⚡ WPS Interface Overview**
29
29
30
30
The Albatross service exposes a single core process, typically named **Drought**, which requires minimal, intuitive inputs to perform complex seasonal forecasts.
31
31
@@ -38,23 +38,22 @@ The service is designed to work with **only four mandatory user inputs** to defi
38
38
3. **Target Season:** Define the season of interest (e.g., JFM, DJF).
39
39
4. **Phase mode:** Define the level of El Niño/La Niña phase analysis.
40
40
41
-
---
42
41
43
-
### 🚀 Operational Forecasting Modes
42
+
**🚀 Operational Forecasting Modes**
44
43
45
44
Albatross supports two distinct operational modes, governed by the ``End year`` parameter, allowing for seamless transition from model verification to real-time prediction.
46
45
47
46
48
-
49
-
#### 🗓 Hindcast (Model Calibration and Verification)
47
+
**🗓 Hindcast (Model Calibration and Verification)**
50
48
51
49
This mode is used for **training, calibrating, and verifying** the predictive skill of the NIPA model using historical data.
52
50
53
51
* **Goal:** To assess how well the model *would have* performed historically.
54
52
* **Availability:** Set the ``End year`` parameter to a year **prior to the current year** (e.g., 2016).
55
53
* **Output:** A time series of historical hindcast predictions, suitable for calculating model skill scores (R-squared, correlation, RMSE, etc.).
56
54
57
-
#### 🔮 Operational Forecast (New Feature!)
55
+
56
+
**🔮 Operational Forecast (New Feature!)**
58
57
59
58
This new mode enables the generation of predictions for the current or upcoming season using the most recent available climate index data.
60
59
@@ -63,9 +62,8 @@ This new mode enables the generation of predictions for the current or upcoming
63
62
* **Process:** The model is trained on historical data, and then uses the current month's teleconnection index value (e.g., the latest available ONI) to generate the forecast value.
64
63
* **Output:** A single point prediction for the specified target season and year.
65
64
66
-
---
67
65
68
-
### 🔧 Detailed Input Requirements
66
+
**🔧 Detailed Input Requirements**
69
67
70
68
To run the WPS, you must provide the following parameters:
71
69
@@ -95,8 +93,6 @@ To run the WPS, you must provide the following parameters:
95
93
| Phase mode (int) | Whether to apply NIPA separately for El Niño and La Niña phases (1 or 2) |
0 commit comments