FIRED RUSSIA

This is event- and daily-level polygons for the Fire event delineation (FIRED) product for RUSSIA from November 2001 to July 2021. It is derived from the MODIS MCD64A1 burned area product (see https://lpdaac.usgs.gov/products/mcd64a1v006/ for more details). The MCD64A1 is a monthly raster grid of estimated burned dates. Firedpy (www.github.com/earthlab/firedpy) is an algorithm that converts these rasters into events by stacking the entire time series into a spatial-temporal data cube, then uses an algorithm to assign event identification numbers to pixels that fit into the same 3-dimensional spatial temporal window. This particular dataset was created using a spatial parameter of 1 pixel and 5 days. The primary benefit to this dataset over others is the ability to calculate fire spread rate. For each of these products (events and daily) the event identification numbers are the same, but the event-level product has only single polygons for each entire event, while the daily product has separate polygons for each date per event.本数据集为2001年11月至2021年7月期间俄罗斯区域的火灾事件划定（Fire Event Delineation, FIRED）产品的事件级与日级矢量多边形数据。该数据集源于MODIS MCD64A1过火面积产品（详细信息请参阅：https://lpdaac.usgs.gov/products/mcd64a1v006/）。MCD64A1产品为逐月生成的过火日期估算栅格网格。Firedpy算法（访问地址：www.github.com/earthlab/firedpy）可将上述栅格数据转换为火灾事件：首先将全时序数据堆叠为时空数据立方体，随后通过算法为落入同一三维时空窗口内的像元分配统一的事件标识编号。本数据集采用1个像元、5天的空间参数生成。相较于其他同类数据集，本数据集的核心优势在于可计算火灾蔓延速率。两类产品（事件级与日级）的事件标识编号保持一致，但事件级产品仅为每个完整火灾事件生成单个矢量多边形，而日级产品则会为每个事件的单日过火范围分别生成独立矢量多边形。