rajvivan/gcc-synthetic-temperature-2022-2024

GCC Synthetic Temperature Dataset (2022–2024)

8 stations across 6 GCC countries, 3 climate zones — 210,432 hourly observations

🔗 Links

• Dataset: https://huggingface.co/datasets/rajvivan/gcc-synthetic-temperature-2022-2024
• Main Project: https://huggingface.co/rajvivan/thermal-risk-aware-dc-orchestration

📊 Stations

#	Station	Country	Zone	T_max	RH Range
1	Dammam	Saudi Arabia	Coastal humid	49.2°C	15–95%
2	Kuwait City	Kuwait	Coastal humid	50.1°C	8–85%
3	Manama	Bahrain	Coastal humid	46.8°C	25–92%
4	Doha	Qatar	Coastal humid	48.5°C	18–88%
5	Muscat	Oman	Mountain fringe	48.0°C	10–78%
6	Al Ain	UAE Interior	Hyper-arid	49.8°C	1–60%
7	Ha'il	Saudi Arabia	Hyper-arid	47.5°C	3–40%
8	Salalah	Oman	Monsoon fringe	38.0°C	40–95%

🔬 Generation Model

T(t) = T_base + A·sin(2π(t-φ)/24) + H·G(t;μ,σ²) + W·S(t) + N(0,σ²_noise)
RH(t) = RH_base + RH_amp·(coastal_penalty·exp(-d²/2σ²) + seasonal·sin(2π(t-φ_RH)/24))

Where:

• G(t): Multi-day Gaussian heatwaves with realistic onset, duration (24–120h), and intensity (3–12°C)
• S(t): Shamal wind temperature drops (5–10°C over 2–4h, 8–15 events per summer)
• coastal_penalty: Gaussian decay with distance from Arabian Gulf coast
• Seasonal: 2°C annual amplitude for summer-winter variation

Zone-Specific Features

Coastal Humid Belt (Dammam, Kuwait City, Manama, Doha):

• Strong humidity: 70–95% summer mornings from warm shallow Gulf waters
• Frequent shamal events (4–5% daily probability in summer)
• Moderate diurnal swing (8–12°C)

Hyper-Arid Interior (Al Ain, Ha'il):

• Extreme temperatures: 48–50°C peaks, <5% humidity
• Massive diurnal swing: 15–22°C
• Rare shamal events, strongest heatwaves (up to 12°C intensity)

Mountain/Monsoon Fringe (Muscat, Salalah):

• Muscat: orographic afternoon inversions, moderate humidity
• Salalah: June–September khareef season — temperature drops 5°C, humidity jumps 30%

📈 Usage

import pandas as pd

# Load GCC station data
df = pd.read_csv('gcc_temperature_data/dammam_hourly_2022_2024.csv', 
                  index_col='time', parse_dates=True)
print(df.head())
#               temperature_2m  relativehumidity_2m
# time
# 2022-01-01 00:00:00         16.2                  62.0
# 2022-01-01 01:00:00         15.8                  65.3

Integration with V5 TRWO Environment

from v5_env import V5Env, V5Config, load_trace

# Load GCC synthetic data
gcc_temp = pd.read_csv('gcc_temperature_data/kuwait_city_hourly_2022_2024.csv',
                        index_col='time', parse_dates=True)

# Create env with GCC temperature trace
env = V5Env(
    config=V5Config(num_racks=10),
    real_temp_trace=gcc_temp['temperature_2m'].values,
    real_humidity_trace=gcc_temp['relativehumidity_2m'].values / 100.0,
)

🏭 DC-Expert Context

These stations were selected to cover the full range of GCC data center locations:

1. Kuwait City / Dammam: Major hyperscale deployments (Google, AWS regions)
2. Dubai / Al Ain: Moro Hub, Khazna, Equinix DX1/DX2
3. Doha: Meeza, Ooredoo DCs
4. Muscat: Omantel, Equinix MC1
5. Salalah: Emerging edge DC market (coolest GCC site)

The coastal humid stations are particularly important because high humidity degrades air-side economizer effectiveness and increases wet-bulb temperature, reducing condenser efficiency — a 5–15% additional COP penalty beyond temperature alone.

📝 Citation

@dataset{gcc-synthetic-temperature-2022-2024,
  author = {Raj Vivaan},
  title = {GCC Synthetic Temperature Dataset for Data Center Thermal Research},
  year = {2026},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/rajvivan/gcc-synthetic-temperature-2022-2024}
}

📜 License

Apache 2.0 — use freely for research and commercial applications.