Requirements & Dependencies¶

Overview¶

This document provides comprehensive details about all hardware requirements, software dependencies, and optional components for running EEMT workflows at various scales.

Hardware Requirements¶

Minimum Specifications¶

Performance Scaling¶

By Dataset Size¶

• Small (< 100 km²): 4 cores, 8 GB RAM
• Medium (100-1000 km²): 8 cores, 16 GB RAM
• Large (1000-10000 km²): 16 cores, 32 GB RAM
• Continental (> 10000 km²): 32+ cores, 64+ GB RAM, distributed computing

By Resolution¶

• 30m DEM: Base requirements
• 10m DEM: 2x memory requirement
• 1m LiDAR: 10x memory requirement, tiling recommended

Software Dependencies¶

Core Geospatial Stack¶

Python Environment¶

Workflow Management¶

Web Interface Dependencies¶

Operating System Support¶

Linux (Primary Platform)¶

Supported Distributions: - Ubuntu 20.04 LTS, 22.04 LTS, 24.04 LTS - Debian 10 (Buster), 11 (Bullseye), 12 (Bookworm) - CentOS 7, CentOS Stream 8/9 - RHEL 7, 8, 9 - Rocky Linux 8, 9 - AlmaLinux 8, 9 - Fedora 36+

Package Managers:

# APT (Ubuntu/Debian)
sudo apt update && sudo apt install [packages]

# YUM/DNF (RHEL/CentOS/Fedora)
sudo yum install [packages]  # or
sudo dnf install [packages]

macOS¶

Supported Versions: - macOS 11 (Big Sur) - macOS 12 (Monterey) - macOS 13 (Ventura) - macOS 14 (Sonoma)

Architecture Support: - Intel x86_64 - Apple Silicon (M1/M2/M3) with Rosetta 2

Package Manager:

# Homebrew installation
brew install grass gdal python@3.11

Windows¶

Supported Versions: - Windows 10 (version 2004+) - Windows 11 - Windows Server 2019, 2022

Installation Methods: 1. WSL2 (Recommended):

wsl --install
# Then follow Linux instructions

1. Docker Desktop:
2. Native Windows containers

3. WSL2 backend for Linux containers

4. Native (Limited support):

5. OSGeo4W installer
6. Conda/Mamba environments

Optional Components¶

Performance Enhancements¶

Development Tools¶

Monitoring & Debugging¶

Container Requirements¶

Docker Installation¶

Minimum Docker Version: 20.10+

# Check Docker version
docker --version

# Required features
- Buildkit support
- Multi-stage builds
- Volume mounts
- Network creation

Container Resources¶

Default Limits:

# docker-compose.yml
services:
  eemt-web:
    deploy:
      resources:
        limits:
          cpus: '4'
          memory: 8G
        reservations:
          cpus: '2'
          memory: 4G

Storage Requirements¶

Container Images: - Base Ubuntu: ~500 MB - EEMT with dependencies: ~2.5 GB - With cached climate data: ~10 GB

Volume Mounts: - Input data: /data/input - Results: /data/output - Cache: /data/cache

Network Requirements¶

Internet Connectivity¶

Required for: - Climate data downloads (DAYMET) - Container image pulls - Package installations - Updates

Bandwidth Requirements: - Minimum: 10 Mbps for basic operations - Recommended: 100+ Mbps for large datasets - DAYMET downloads: ~1 GB per year of data

Firewall Ports¶

Cloud Platform Requirements¶

AWS¶

EC2 Instance Types: - Development: t3.large (2 vCPU, 8 GB) - Production: c5.4xlarge (16 vCPU, 32 GB) - GPU-enabled: p3.2xlarge (8 vCPU, 61 GB, V100 GPU)

Storage: - EBS: gp3 volumes, 100+ GB - S3: For input/output data

Google Cloud¶

Compute Engine: - Development: e2-standard-4 - Production: c2-standard-16 - GPU-enabled: n1-standard-8 with T4 GPU

Azure¶

Virtual Machines: - Development: Standard_D4s_v3 - Production: Standard_F16s_v2 - GPU-enabled: Standard_NC6s_v3

HPC Systems¶

SLURM Requirements:

#SBATCH --nodes=1
#SBATCH --ntasks=16
#SBATCH --mem=32GB
#SBATCH --time=24:00:00
#SBATCH --partition=standard

Modules:

module load grass/8.4
module load gdal/3.8
module load python/3.11

Verification Commands¶

Check All Dependencies¶

#!/bin/bash
# Save as check_requirements.sh

echo "Checking EEMT Requirements..."
echo "=============================="

# Python
echo -n "Python: "
python3 --version 2>/dev/null || echo "NOT FOUND"

# GRASS GIS
echo -n "GRASS GIS: "
grass --version 2>/dev/null | head -1 || echo "NOT FOUND"

# GDAL
echo -n "GDAL: "
gdalinfo --version 2>/dev/null || echo "NOT FOUND"

# Docker
echo -n "Docker: "
docker --version 2>/dev/null || echo "NOT FOUND"

# CCTools
echo -n "Makeflow: "
makeflow --version 2>/dev/null | head -1 || echo "NOT FOUND"

# Python packages
echo -e "\nPython Packages:"
python3 -c "
import importlib
packages = ['numpy', 'pandas', 'rasterio', 'xarray', 'geopandas']
for pkg in packages:
    try:
        mod = importlib.import_module(pkg)
        print(f'  {pkg}: {mod.__version__}')
    except ImportError:
        print(f'  {pkg}: NOT INSTALLED')
" 2>/dev/null

# System resources
echo -e "\nSystem Resources:"
echo "  CPU Cores: $(nproc 2>/dev/null || sysctl -n hw.ncpu 2>/dev/null || echo 'Unknown')"
echo "  Total RAM: $(free -h 2>/dev/null | awk '/^Mem:/{print $2}' || echo 'Unknown')"
echo "  Available Disk: $(df -h . | awk 'NR==2{print $4}')"

Troubleshooting Dependencies¶

Version Conflicts¶

# Check for conflicts
pip check

# Force reinstall with specific versions
pip install --force-reinstall numpy==1.26.0 pandas==2.1.0

Missing Libraries¶

# Find missing libraries (Linux)
ldd $(which grass) | grep "not found"

# Install missing libraries
sudo apt install libgdal-dev libproj-dev  # Ubuntu/Debian
sudo yum install gdal-devel proj-devel     # CentOS/RHEL

Permission Issues¶

# Fix permission problems
sudo chown -R $USER:$USER ~/.grass8
chmod 755 ~/grassdata
chmod 600 ~/.eemt-makeflow-password

For installation instructions, see the Installation Overview.