How to pass chunks of StructedBuffers into a compute shader

I have a sparse voxel octree that is flatten for the gpu now I have the problem that this only allows one octree at a time and not multiple chunks, I thought about combining the octrees but that would mean I have to store the old octree per chunk and the combined octrees (which would double the memory) maybe there is a better solution I am not aware of yet.

StructuredBuffer<GPUOctreeNode> _SVOBuffer;
StructuredBuffer <float4> _SVOData;

The GPUOctreeNode looks like this, it points to the _SVOData with the dataIndex

struct GPUOctreeNode {
    float3 position;
    float halfSize;
    int childIndex;
    int dataIndex;
};

or I would store the octree chunks in a combined structuredBuffer only and use compute shaders to add and remove chunks during runtime, this way I dont need to keep the old per chunk octree in memory